Designations for the 2006 PM2.5 Standards: Evaluating the Nine Factors in Setting Nonattainment Area Boundaries
Part 2 – Conceptual Model for Evaluating High PM2.5 Days and its Influencing Emission Sources
Neil Frank
EPA Office of Air Quality Planning and Standards
For Presentation at EPA State / Local / Tribal Training Workshop: PM 2.5 Final Rule Implementation and 2006 PM 2.5 Designation Process June 20-21, 2007
1
The 9 Designation Factors
To Help Determine Nearby Area of Influence for 24-hr NAAQS Violations
Population and Urbanization Emissions
Traffic & Commuting
Air Quality
Non Attainment Boundaries
Growth
Meteorology Political and Other Boundaries
Current Emission Controls
Topography
Air Quality is one of the most Important Designation Factors
2
Topics to be Covered
• • • • Conceptual model for high PM days Seasons when exceedances occur Composition of the high days Analytical tools
– SLICE technique - for evaluating urban contributions to high days – Residence time analysis – for assessing nearby contributing source regions using back trajectories and emissions data – Gradient analysis – for identifying days with potential high source-oriented impacts
3
Conceptual Model for High PM2.5 Days
• How to define high PM2.5 days? • What is the typical “daily increment” for high PM days in relation to the annual average? • What is the urban contribution above regional levels?
4
Conceptual Model for High PM2.5 Days
What high PM2.5 days to consider?
• “High PM2.5 Days” Associated with the 98th percentile
Not just one day per year Select all candidate days e.g. top 5% or days > 30 - 35ug/m3 Summarize by season to distinguish varying conditions
5
Conceptual Model for High PM2.5 Days
High Daily PM2.5 has Urban and Regional Components
40ug/m3
Typical Daily Increment - Example
24 16
Example
PM2.5 High Day Value Seasonal Average
+
Typical Daily Increment
• The annual average PM2.5 (urban background) is the stuff that is there on a day-today basis. –Comes from nearby and more distant areas –Can be estimated by seasonal average PM2.5 concentration of non-high days –Includes contributions from all nearby surrounding counties –Can be estimated using the traditional urban increment approach • The daily increment (on top of annual average urban background) also has regional and local contributions. - Key issue: what counties and sources from the urban area contribute to the typical daily increment?
6
Conceptual Model for High PM2.5 Days
An approach to partition typical levels into urban and regional components
•
•
–
Urban Increment Analyses as used in 2004/2005 PM2.5 Designations
Urban sources in the Eastern US contribute at least 4-6 ug/m3 to annual average PM2.5
Probably even larger urban contribution in western US cities
•
–
Carbon is significant component of average PM2.5 mass, but metro area emissions typically are much less than SO2 and NOx
Weighted emissions score developed to give additional weight to nearby direct carbon emissions as they contribute to the urban background
7
Air Quality - Annual Average PM2.5 Conceptual Diagram PM [µg/m³]
Annual Average PM2.5
40 35 30
25
20 15
Urban areas ~14-20 ug/m3
Countryside ~10-12 ug/m3
10
urban increment regional contribution
natural background
8
Western US urban areas may have smaller regional contribution
Air Quality - High Daily PM2.5 Concentrations Conceptual Diagram PM [µg/m³]
High Daily PM2.5 Concentrations
40 35
Urban areas ~30-40 ug/m3
urban increment
Countryside ~18-30 ug/m3
Larger “urban island” on peak days
30
25
20 15
regional contribution
10
natural background Focus of new analyses: understanding what emissions contribute to urban increment
Western US urban areas may have smaller regional contribution
9
Conceptual Model for High PM2.5 Days
Emissions
Population and Urbanization Traffic & Commuting
Source region considerations
•
Air Quality
Air Quality
Growth Non Attainment Boundaries
Meteorology
Current Emission Controls Political and Other Boundaries
Topography
Role of Regional vs Urban vs Micro-scale Influences
– – – On high days particularly in the east, regional emissions often provide a “base” amount of pollution Urban-wide and nearby emissions also contribute significantly to high days: “urban island” effect In some cases, there may be a micro-scale effect from a single source or small group of sources
• Does not help define NA boundaries, unless it is the only contributing source
(Note: “urban” can mean large metropolitan area or smaller city)
10
Conceptual Model for High PM2.5 Days
Seasons when exceedances occur
• Time of Year for Exceedances- varies by Geographic Region – SE: Mostly summer – Industrial Midwest (IMW), Mid-Atlantic, So. CA: Winter and summer – NW, UT, NM, Middle CA: Mostly or exclusively
Winter
11
% 50 50
40
Percent of 2003-05 Days > 35 ug/m3, by Month (NW)
Based on all FRM Site-days throughout the Regional Domain
Percent of 2003-05 Days > 35 ug/m3, by Month (UT)
50
Percent of 2003-05 Days > 35 ug/m3, by Month (IMW)
50
Based on all FRM Site-days throughout the Regional Domain
Based on all FRM Site-days throughout the Regional Domain
NW
40 30
UT
40
IMW
30
Percent of 2003-05 FRM Days > 35 ug/m3 by Month
30
25 20
10
Based on all sites which violate 24-hr NAAQS
20
20
10
10
0
0 J
0
0
J PLOT F M A M J J PvX A S O N D
PLOT
J F M A M J J A S O N D
GT 35 ug/m3 at FRM sites PvX
Based on all FRM Site-days throughout the Regional Domain
F
M
A
M
J
J
A
S
O
N
D
J PLOT
F
M
A
M
J
J PvX
A
S
O
N
D
GT 35 ug/m3 at FRM sites
J F M A M J J A S O N D
GT 35 ug/m3 at FRM sites Based on all FRM Site-days throughout the Regional Domain
Percent of 2003-05 Days > 35 ug/m3, by Month (Mid. Cal)
Percent of 2003-05 Days > 35 ug/m3, by Month (N.Eng-MidAtl)
50
60
50
Mid CA
40
MidAtl
40
30
30
20
20
10
10
0 J PLOT F M A M J J PvX A S O N D
0 J PLOT F M A M J J PvX A S O N D
J F M A M J J A S O N D
GT 35 ug/m3 at FRM sites
J F M A M J J A S O N D
GT 35 ug/m3 at FRM sites
Based on all FRM Site-days throughout the Regional Domain
% 50 50
40
Percent of 2003-05 Days > 35 ug/m3, by Month (S. Cal)
Based on all FRM Site-days throughout the Regional Domain
Percent of 2003-05 Days > 35 ug/m3, by Month (Las Cruces)
50
Based on all FRM Site-days throughout the Regional Domain
Percent of 2003-05 Days > 35 ug/m3, by Month (SE)
50
S.CA
40
La Cruces, NM
40 30
SE
30
30
2520
10
20
20
10
10
12
J F M A M J J A S O N D
0
0 J F M A M J J A S O N D
0 J F M A M J J A S O N D
0
Conceptual Model for High PM2.5 Days
Population and Urbanization Emissions
Composition data are important
1) Composition
Air Quality
Traffic & Commuting
Air Quality
Growth Non Attainment Boundaries
Meteorology
Current Emission Controls Political and Other Boundaries
Topography
– Indicate which sources are contributing to average and high PM2.5 values
• Varies across country
– –
Warm season exceedances: Mostly sulfate +
organic carbon
Cold season exceedances: Nitrate (at higher
latitudes and in Western US) + sulfate + carbon; Carbon may dominate in some locations (e.g. MT, ID) Gaps in speciation data for certain areas
–
13
Air Quality
Composition on Annual Average and
High PM2.5 Days (From PM Staff Paper)
Some source categories and regional influences may be Birm more important for high concentration days
High PM2.5 days have:
S
More Sulfate 14 More Nitrate
•
•
•
•
Comparing average of 5 highest days during 2003, regional sources of sulfates and nitrates are larger contributors to peak day concentrations than to annual average (selected city analysis) Composition can vary from high day to high day Carbon can be smaller as % -- but still larger in absolute concentration values -compared to the average Note: All the new analyses present “FRM” composition with the peer-reviewed “SANDWICH” Technique
– As used in CAIR and PM2.5 RIA
Nitrate TCM This analysis shows Sulfate PM2.5 Composition of the ambient aerosol (not adjusted to represent FRM mass) Crustal
Atlanta
NE
NYC
Cleveland
Chicago
MW
St .Louis
SLC
UT
Fresno
CA
From PM Staff Paper (Rao et al)
% 50 50
40
Percent of 2003-05 Days > 35 ug/m3, by Month (NW)
Based on all FRM Site-days throughout the Regional Domain
Percent of 2003-05 Days > 35 ug/m3, by Month (UT)
50
Based on all FRM Site-days throughout the Regional Domain
Percent of 2003-05 Days > 35 ug/m3, by Month (IMW)
50
Based on all FRM Site-days throughout the Regional Domain
NW
Cold
40
UT
Cold
Avg High 3 Cold
40
IMW
Cold
Avg High 3 Cold
Warm
Avg High 3 Warm
30
or
30
30
2520
10
20
20
10
10
0
0 J F M A M J J A S O N D
0 J F M A M J J A S O N D
0 J F M A M J J A S O N D
“Example” Composition for High Days [“Warm” Season (May-Sept) & “Cold”]
PLOT GT 35 ug/m3 at FRM sites PvX PLOT GT 35 ug/m3 at FRM sites PvX PLOT GT 35 ug/m3 at FRM sites PvX
Percent of 2003-05 Days > 35 ug/m3, by Month (Mid. Cal)
Based on all FRM Site-days throughout the Regional Domain
60
50
Mid CA
Cold
But sites can be different within each “domain” MidAtl
50 40 30
Percent of 2003-05 Days > 35 ug/m3, by Month (N.Eng-MidAtl)
Based on all FRM Site-days throughout the Regional Domain
Cold
Warm
40
or
30
20
20
10
10
0 J PLOT F M A M J J PvX A S O N D
0 J PLOT F M A M J J PvX A S O N D
% 50 50
40
J F M A M J J A S O N D Pies represent average ofof3 highest by Month (Las Cruces) year per season, usingDays > 35 ug/m3, by Month (SE) days per SANDWICH Percent 2003-05 Days > 35 ug/m3, Percent of 2003-05 Percent of 2003-05 Days > 35 ug/m3, by Month (S. Cal)
GT 35 ug/m3 at FRM sites
J F M A M J J A S O N D
Based on all FRM Site-days throughout the Regional Domain
GT 35 ug/m3 at FRM sites
Based on all FRM Site-days throughout the Regional Domain
Based on all FRM Site-days throughout the Regional Domain
50
50
S.CA
Cold
Avg High 3 Cold
Warm
Avg High 3 Warm
40
La Cruces NM
Cold
Avg High 3 Cold
SE
40 30
Warm
Avg High 3 Warm
30
30
2520
10
20
El Paso STN
20
10
10
15
J F M A M J J A S O N D
0
0 J F M A M J J A S O N D
0 J F M A M J J A S O N D
0
Composition is often similar among the high days
IMW - 551330027 Wisconsin Milwaukee-Waukesha,WI 24-hr DV= 36 ug/m3 Annual DV= 13.5 ug/m3
"cold" days > 30 = 5 Total 3 yr obs = 161 "warm" days > 30 = 1
Avg conc. - top 3 values/yr:
39.0 ug/m3
41.1 ug/m3
S.CA Milwaukee, 2003-05 24 Average Cool Season Warm Season
AVG Avg High 3 Cold Avg High 3 Warm
22 20 18
PM2.5 and component mass, ug/m3
3 highest PM2.5 days > 30ug/m3 Per season, Milwaukee, WI (2003-05)
16 14 12 10 8.0 6.0 4.0 2.0 0.0 -2 2002 2003
EC Passive
Sulfate_mass OCMmb
Nitrate_mass Crustal
PM2.5 days > 30ug/m3
Measured PM2.5 mass, ug/m3 2004 2005
Nitrate_mass Crustal
2006
16
EC Passive
Chem Black line is difference between OCMmb and OCM14
Sulfate_mass OCMmb
Conceptual Model for High PM2.5 Days
An approach to partition total daily increment into urban and regional components
IMW 550790026 [Tot warm days= 142] Wisconsin Milwaukee-Waukesha,WI [High days= 3] 550790026 [Tot cold days= 183] Wisconsin Milwaukee-Waukesha,WI [High days= 8] 24-hr DV= . ug/m3 Annual DV= . ug/m3 24-hr DV= . ug/m3 Annual DV= . ug/m3
Cool
Warm
65 60 55
65
PM2.5 and component mass, ug/m3
50 45 40 35 30 25 20 15 10 5.0 0.0 1 2 . 4 3 6 1 2 . 6 4 6 3 2 . 3 4 4 3 2 . 7 0 5
PM2.5 and component mass, ug/m3
Cool Daily composition 55 Warm minus Seasonal avg. season 50 season
45 40 35 30 25 20 15 10 5.0 0.0
60
Next subtract the daily composition from the seasonal average PM2.5 Use resident time weighted emissions to partition each component of total daily Increments into urban & regional contributions (% of RTWE in local area)
AS
SA
3 3 . 9 6 4
3 5 . 0 4 3
3 5 . 7 4 4
4 0 . 0 5 7
4 1 . 0 2 0
4 1 . 5 3 8
Measured PM2.5 mass
1 1 . 8 9 9
1 2 . 4 2 0
3 4 . 8 4 2
4 5 . 6 0 7
4 6 . 6 6 4
under other bars are the daily concentration. --Estimated INCREMENT COMPOSITION is shown
Sulfate_mass Nitrate_mass EC Nitrate_mass can be EC backgroundChem PM2.5 OCMmb Crustal Passive PM2.5 days > 30ug/m3 Crustal Passive 17 estimated using seasonal average PM2.5 Per season & year, Milwaukee, WI (2003-05) First bars are seasonal and annual averages (latter if data isbars are seasonal and annual averages (latter if data is complete) concentration of non-high days First complete) daily concentration. --Estimated INCREMENT COMPOSITION is shown Values under other bars are the
Sulfate_mass The urban OCMmb
Analytical Tools
to help identify boundaries and develop SIPs
• SLICE technique - for evaluating urban contributions to high days • Residence time analysis – for assessing nearby contributing source regions using back trajectories and emissions data • Urban gradient analysis – for identifying whether there are any sites predominantly affected by a single source
18
Analytical Tools
Identify urban PM2.5 and gradients
•
Air Quality
Population and Urbanization Emissions Traffic & Commuting
Air Quality
Growth Non Attainment Boundaries
Meteorology
Current Emission Controls Political and Other Boundaries
Topography
“SLICE” to identify “urban island” days and relative urban amount of PM2.5 mass
• Evidence of urban source contributions
•
Air Quality
Urban “gradient” technique
•
Met
Emissions
Evidence of predominant strong nearby source influence
Daily urban portion of PM2.5
19
Analytical Tools - Residence Time Analysis
Emissions
Population and Urbanization Traffic & Commuting
Where did the air parcel come from on high concentration days?
Air Quality
Air Quality
Growth Non Attainment Boundaries
Meteorology
Current Emission Controls Political and Other Boundaries
Topography
Met 1) Transport patterns producing a potential source region
• •
Use trajectories and “Residence-Time Analysis” to find upwind probability fields. For PM2.5 mass or its components
Focus on the ensemble of “High PM2.5 days”, by season for subsequent linking to composition pattern.
Days with identified “urban islands” are more important
•
Local pollution roses (annual vs. high days) would also be helpful to identify nearby sources.
Residence time probability plots with HYSPLIT trajectories have been used by Kinski, Poirot and others to identify potential source regions.
20
Analytical Tools- Residence time weighted emissions
What are the most likely contributing emissions?
Emissions
Population and Urbanization Emissions Traffic & Commuting
Air Quality
Growth Non Attainment Boundaries
Meteorology
Current Emission Controls Political and Other Boundaries
Topography
1)
Spatial distribution of emissions by season
• • • • Developed from monthly emissions for precursors and direct PM: (SO2, NOx, Carbon, Crustal ) The importance of each precursor pollutant can be guided by the composition of the high PM2.5 day. consider monthly emissions corresponding to the affected PM component according to typical composition by season. Some precursors will not be considered or could be downweighted. e.g. crustal (year-round) and NOX (summer).
Met
•
Residence time weighted emissions
• Use probability that air parcel passed over an area to weight emissions as potential contributors to the high day concentration impacts High probability nearby contributing emissions can be identified for each PM2.5 contributor
21
Emissions
Air Quality
•
Summary
• Identifying the area of emission influence considers contributions for
– each “high PM2.5 day” and – urban average background on top of which are the daily impacts
• High concentration days with evidence of urban influence (i.e. with urban islands) are more important
– The magnitude of urban island can help define the daily urban contributions.
• In combination with daily and average speciation data, by season of the year
– Emissions with high probability of trajectory residence time are important to assess high day impacts. – Average emissions and typical wind patterns help understand the sources contributing to the urban “background” – Both used to understand the relative importance of the various nearby contributing emissions (e.g. direct PM vs SO2 vs NOx).
22