Single Source Ozone and PM Modeling

Ralph Morris ENVIRON International Corporation

rmorris@environcorp.com

EPA’s 9th Conference on Air Quality ModelingNon-Guideline Applications

October 10, 2008

EPA Guidelines/Guidance for Air Quality Modeling Assessments

AERMOD for near-source (< 50 km) air quality impacts of primary pollutants (EPA, 2005)CALPUFF for far-field (> 50 km) air quality impacts of primary pollutants (EPA, 2003)– CALPUFF adopted for SO2 and primary PM PSD

increments and NAAQSs, not AQRV analysis

Photochemical Grid Models (PGMs) recommended for ozone and secondary components of PM (e.g., SO4 & NO3) (EPA, 2007)– Lagrangian models (e.g., AERMOD and CALPUFF) only

suitable for primary components

New Ozone and PM2.5 NAAQS

New 0.075 ppm 8-hour and 35 µg/m3 24-hr PM2.5NAAQSs will bring many more areas into nonattainment– PM2.5 NAAQS increases importance of secondary PM2.5

Capability needed to obtain individual contributions to ozone and PM2.5 concentrations, deposition and visibility

Current guideline models have no (AERMOD) or highly simplified (CALPUFF) representation of chemistry

Photochemical Grid Models (PGMs) have capability to correctly treat chemistry – But how can they resolve and correctly simulate near source

plume chemistry and dispersion?

Why PGMs Have Not Been Used to Address “Single Source” Impacts

PGMs can only resolve impacts to the grid resolution– Fine grid size needed near the source to resolve near-

source plume chemistry and dispersionNeed many grid cells to assess downwind impactsHigh computer resource requirements

Must account for all emission sourcesNeeded to correctly simulate chemistry

Databases more costly to developMM5/WRF applicationsSMOKE or other emissions model

More expertise needed in their application

Recent Developments for “Single Source” Modeling using PGMs

Two-way interactive grid nesting– Allows fine grid over sources with coarser grid downwind

when plumes are larger

Flexi-nesting– Can specify fine grid to resolve point source plume chemistry

and dispersion without providing met and emission inputs

Full Chemistry Plume-in-Grid Modules– Treats unique near-source chemistry of point source plumes

PM and Ozone Source Apportionment– Allows individual source(s) assessments

Computational advancesAvailability of PGM Databases and model set ups– RPOs, AIRPACT, SIPs, etc.

Advanced Grid Nesting

Two-way interactive grid nesting– Allows specification of high resolution grid over sources

with coarser grids downwind where plumes are larger

Flexi-Nesting– Interpolate meteorology, emissions and/or other inputs

for nested fine grid from coarse grid data– Allows fine grid treatment of point source plumes

Available within the CAMx model (just specify where fine grid domains are desired in job script)

Have developed tool to generate flexi-nest fine grid inputs for CMAQ (for EPA/OAQPS)

Full Chemistry Plume-in-GridIncremental chemistry approach allows full gas-phase, aqueous-phase and aerosol chemistry within Plume-in-Grid modules– CAMx Incremental Reactions for Organics and NOx

(IRON) Plume-in-Grid (PiG) treatment– CMAQ Advanced Plume Treatment (APT)

• Calculate chemical kinetic reaction rates using total concentrations (PiG + Grid)

• Apply chemical rates to incremental concentrations within puffs (PiG alone)

• When size of plume is commensurate with grid cell size release incremental PiG concentrations to grid model

Full Chemistry Plume-in-Grid

Important to simulate proper chemistry in early evolution of point source plumes

Very little if any ozone and secondary PM2.5formed under Stage 1 and 2 plume chemistry conditions for large NOx sources

Stack

PM and Ozone Source Apportionment

PM and Ozone Source Apportionment Technology (OSAT/PSAT) available in CAMx– Uses reactive tracers that operate in parallel to host

model and tracks ozone and PM formation back to emission source regions and categories

– PSAT has five families of tracers: SO4; NO3/NH4; Primary PM; SOA; and Hg

Similar approaches in CMAQ (TSSA and PPTM)Allows for identification of the ozone and PM impacts from several individual sources or groups of sources in single run

Example “Single Source” PGM Applications

Texas “Group BART” application– CAMx 36/12 km with PiG and PSAT

Estimation of individual contributions of 31 point sources to annual PM2.5 in the eastern U.S.– Individual point source contributions to 2009 annual

PM2.5 concentrations– Visibility Improvements for States and Tribal Association

of the Southeast (VISTAS) and Association for Integrated Planning of the Southeast (ASIP)

Annual PM2.5 SIP modeling for St. Louis– Effects of local sources on PM2.5 nonattainment

Texas Group BART AnalysisCENRAP 2002 36 km modeling CAMx database– Add 12 km flexi-nest grid covering Texas and nearby Class I areas– Use IRON PiG for Texas BART Source– Use PSAT to obtain PM2.5 contributions of groups of Texas BART

sources for comparison with 0.5 deciview threshold

-2736 -1872 -1008 -144 720 1584 2448-2088

-1656

-1224

-792

-360

72

504

936

1368

1800

km

km

-936 -816 -696 -576 -456 -336 -216 -96 24 144 264 384

-1620

-1500

-1380

-1260

-1140

-1020

-900

-780

-660

-540

-420

km

km

VISTAS/ASIP 36/12 km Domains

2002 Annual Runs– 4 Quarters w/ ~15 day spin up– MPI w/ 6 CPUs

36 km: 148 x 112 (4 days)12 km: 168 x 177 (10 days)19 Vertical Layers CMAQ V4.51 w/ SOAmods– M3Dry– CBM-IV/AE4/SORGAM– SOAmods: In 2005 VISTAS

enhanced CMAQ to include SOA from sesquiterpenes and isoprene (Morris et al., 2006)

ASIP PM2.5 Point Source ContributionsSome ASIP/VISTAS states wanted to know individual contributions of several point sources to 2009 PM2.5 levels– 31 individual point sources in 6 states identified– Contributions due to SO2 and primary PM emissions requested

CALPUFF considered for assessment– Not consistent with CMAQ full-science chemistry– Provide inconsistent source contributions with 2009 PM2.5 SIP projections

ASIP 36/12 km database inappropriate for individual point sourcemodeling– 12 km grid cell size too coarse to treat chemistry and dispersion of point

source plumes– Use of high enough resolution to resolve point source plume would be

computationally prohibitive– Would need to perform base case and 31 zero-out runs to get individual

source contributionsElected to develop a new CAMx 2002 database:– 12/4 km domain with two-way nested grids– Plume-in-Grid to address near-source chemistry and dispersion– PM Source Apportionment Technology (PSAT) to obtain individual source

contributions

CAMx 12/4 km domain nested within ASIP 12 km CMAQ domain (one-way nesting)CAMx 12/4 modeling using two-way interactive grid nesting2002 base case using standard model2009 base case with PSAT PM2.5 source apportionment for 31 point sources

120 240 360 480 600 720 840 960 1080 1200 1320 1440 1560 1680 1800 1920 2040

-1560

-1440

-1320

-1200

-1080

-960

-840

-720

-600

-480

-360

-240

-120

0

120

240

360

480

VISTAS CMAQ 12 km: 168 x 177 (108, -1620) to (2124, 504)CAMx 12km: 54 x 71 (828, -612) to (1476, 240)

VISTAS CMAQ 12km

CAMx 12km

900 1000 1100 1200 1300 1400

-600

-500

-400

-300

-200

-100

0

100

200

CAMx 12km: 54 x 71 (828, -612) to (1476, 240)

CAMx 04km (includes buffer cells):Steubenville/Wierton 26 x 35 (1316, 80) to (1420, 220)Charleston 50 x 38 (1172, -148) to (1372, 4)Louisville 17 x 26 ( 944, -184) to (1012, -80)Chattanooga/Knoxville 47 x 44 (1028, -496) to (1216, -320)

CAMx 12km

4km Chat

4 km Char

4 km Steu

4 km Loui

CAMx 12/4 DomainsFour 4 km domains:– Charleston-Huntington,

KY/OH/WV– Wheeling-Weirton,

OH/PA/WV– Louisville, IN/KY– Knoxville-Chattanooga,

GA/KY/NC

PSAT to obtain individual PM2.5 contributions from 31 point sourcesPlume-in-Grid for 31 plus other large point sources

Huntington-Ashland and Charleston 4 km Domain Map with FRM sites & Source Locations

1180 1200 1220 1240 1260 1280 1300 1320 1340 1360

-140

-120

-100

-80

-60

-40

-20

0

Marathon Ashland Pet LLC

Ak Steel CorpE I Dupont Inc

Kentucky Power Co Big Sandy Plant

Ohio Valley Electric Corp. Kyger CreekGavin Power Plant

Swva, Inc.

Dupont - Belle

Bayer Cropscience

Appalachian Power - Mountaineer Plant

Flexsys - Nitro PlantAppalachian Power - John E Amos Plant

KY-Boyd-0017

KY-Carter-0500

OH-Lawrence-0010

OH-Scioto-0013

WV-Cabell-0006WV-Kanawha-0010WV-Kanawha-1005

Appalachian Power - Kanawha River Plant

Appalachian Power - Philip Sporn Plant

Huntington AlloysFmc Corporation - Steam PlantUnion Carbide (Dow) So. Charleston PlantClearon Corp.

Huntington-Ashland-Charleston: 4 km Grid; All Sources and SO4+Prim-PM

ASIP 2009 Q1 Projected SO4+EC+POC

0

1

2

3

4

5

6

7

8

9

10

KY-Boyd

-0017

KY-Cart

er-0500

OH-Lawren

ce-00

10

OH-Scio

to-001

3

WV-Cab

ell-000

6WV-K

anawha

-0010

WV-Kan

awha-10

05

Sources

SO4+

EC+P

OC

ug/

m3

BCICOthers

Appalachian Pow er Co.-John E AFlexsys - Nitro PlantAppalachian Pow er Co.-MountainAppalachian Pow er Co.-Philip S

Ohio Pow er - Kammer PlantOhio Pow er - Mitchell PlantPpg Industries, Inc.Clearon Corp.

Bayer CropscienceAppalachian Pow er - Kanaw ha RiUnion Carbide (Dow ) So. CharleFmc Corporation - Steam Plant

Dupont - BelleIsg WeirtonSw va, Inc.Huntington Alloys - A Special

Bow ater New sprint & DirectoryE.U.Du Pont De Nemours And ComPa Pow er Co/Bruce Mansfield PlW.H. Sammis Plant

Wheeling-Pittsburgh Steel CorpCardinal Pow er Plant (CardinalGavin Pow er PlantOhio Valley Electric Corp Kyge

R.E. Burger PlantKentucky Pow er Co Big Sandy PE I Dupont IncAk Steel Corp

Marathon Ashland Pet LlcPsi Energy - Gallagher

ASIP 2009 Annual Projected SO4+EC+POC

0

0.5

1

1.5

2

2.5

3

KY-Boyd

-0017

KY-Cart

er-0500

OH-Lawren

ce-00

10

OH-Scio

to-001

3

WV-Cab

ell-000

6WV-K

anawha

-0010

WV-Kan

awha-10

05

Sources

SO

4+E

C+P

OC

ug/m

3

Appalachian Pow er Co.-John E A

Flexsys - Nitro Plant

Appalachian Pow er Co.-Mountain

Appalachian Pow er Co.-Philip S

Ohio Pow er - Kammer Plant

Ohio Pow er - Mitchell Plant

Ppg Industries, Inc.

Clearon Corp.

Bayer Cropscience

Appalachian Pow er - Kanaw ha Ri

Union Carbide (Dow ) So. Charle

Fmc Corporation - Steam Plant

Dupont - Belle

Isg Weirton

Sw va, Inc.

Huntington Alloys - A Special

Bow ater New sprint & Directory

E.U.Du Pont De Nemours And Com

Pa Pow er Co/Bruce Mansfield Pl

W.H. Sammis Plant

Wheeling-Pittsburgh Steel Corp

Cardinal Pow er Plant (Cardinal

Gavin Pow er Plant

Ohio Valley Electric Corp Kyge

R.E. Burger Plant

Kentucky Pow er Co Big Sandy P

E I Dupont Inc

Ak Steel Corp

Marathon Ashland Pet Llc

Psi Energy - Gallagher

Huntington-Ashland and Charleston:4 km Grid; 31 Facilities

Ak Steel Corp

Swva Inc.

(12.9)

(10.3)

(13.7)

(15.0)

(14.5)

(13.3) (15.0)

(CAMx 4 km 2009 PM2.5 DVs)

ASIP 2009 Annual Projected PSO4

0

0.2

0.4

0.6

0.8

1

1.2

1.4

KY-Boy

d-0017

KY-Cart

er-05

00

OH-Lawren

ce-001

0

OH-Scio

to-0013

WV-Cab

ell-00

06

WV-Kan

awha-0

010

WV-Kan

awha

-1005

Sources

PSO

4 ug

/m3

Huntington- Ashland and Charleston: 31 Facilities; SO4 (left) and Primary PM (right)

ASIP 2009 Annual Projected PEC+POA

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

KY-Boyd

-0017

KY-Carte

r-050

0

OH-Lawrenc

e-0010

OH-Scio

to-00

13

WV-C

abell

-0006

WV-Kan

awha

-0010

WV-Kan

awha

-1005

Sources

PEC

+PO

A u

g/m

3

SO4 Primary PM

Wheeling 4 km Domain Map with FRM sites and Source Locations

1320 1340 1360 1380 1400 142080

100

120

140

160

180

200

220

R.E. Burger Plant

Cardinal Power Plant (Cardinal Operating

Wheeling-Pittsburgh Steel Corp

W.H. Sammis Plant

Pa Power Co/Bruce Mansfield Plt

Isg Weirton

Ppg Industries, Inc.

Ohio Power - Mitchell PlantOhio Power - Kammer Plant

WV-Brooke-0005

WV-Hancock-0011WV-Hancock-1004

WV-Marshall-1002

WV-Ohio-0010

CAMx 04km (includes buffer cells):Steubenville/Wierton 26 x 35 (1316, 80) to (1420, 220)

ASIP 2009 Annual Projected SO4+EC+POC

0

0.5

1

1.5

2

2.5

WV-Brook

e-0005

WV-Han

cock

-0011

WV-Han

cock

-1004

WV-Mars

hall-1

002

WV-Ohio-

0010

Sources

SO4+

EC+P

OC

ug/

m3

Appalachian Pow er Co.-John E A

Flexsys - Nitro Plant

Appalachian Pow er Co.-Mountain

Appalachian Pow er Co.-Philip S

Ohio Pow er - Kammer Plant

Ohio Pow er - Mitchell Plant

Ppg Industries, Inc.

Clearon Corp.

Bayer Cropscience

Appalachian Pow er - Kanaw ha Ri

Union Carbide (Dow ) So. Charle

Fmc Corporation - Steam Plant

Dupont - Belle

Isg Weirton

Sw va, Inc.

Huntington Alloys - A Special

Bow ater New sprint & Directory

E.U.Du Pont De Nemours And Com

Pa Pow er Co/Bruce Mansfield Pl

W.H. Sammis Plant

Wheeling-Pittsburgh Steel Corp

Cardinal Pow er Plant (Cardinal

Gavin Pow er Plant

Ohio Valley Electric Corp Kyge

R.E. Burger Plant

Kentucky Pow er Co Big Sandy P

E I Dupont Inc

Ak Steel Corp

Marathon Ashland Pet Llc

Psi Energy - Gallagher

Wheeling and Steubenville-Weirton: 4 km Grid; 31 Facilities and SO4+Prim-PM

(13.3)

(12.5)

(13.6)

(13.5) (12.8)

Isg Weirton

W.H. Sammis

Wheeling Pittsburgh Steel

Wheeling-Weirton: 31 Facilities; SO4 (left) and Primary PM (right)

ASIP 2009 Annual Projected PSO4

0

0.3

0.6

0.9

1.2

1.5

WV-Brook

e-0005

WV-Hanc

ock-00

11

WV-Hanc

ock-10

04

WV-Mars

hall-1

002

WV-Ohio

-0010

Sources

PSO

4 ug

/m3 ASIP 2009 Annual Projected PEC+POA

0.00

0.30

0.60

0.90

1.20

1.50

WV-Broo

ke-00

05

WV-Han

cock-

0011

WV-Han

cock-

1004

WV-Mars

hall-1

002

WV-Ohio

-0010

Sources

PEC

+PO

A u

g/m

3

SO4

Primary PM

St. Louis 2003-2005 Annual PM2.5Design Values

-500 0 500 1000 1500

-1500

-1000

-500

0

500

Domain 1 (68x68)

Domain 2 (128x149)

Domain 3 (92x113)

St. Louis Regional 36/12

km gridCMAQ V4.5 SOAmods

Projected 2009 and 2012 PM2.5 Design Values at Granite City and East St.

Louis still exceed the annual PM2.5

NAAQS

Evidence that local sources contribute to PM2.5 nonattainment at Granite City Monitor (B) and Washington St. Monitor (A)

Hybrid CAMx 2002 12/4/1 km Modeling with PiG for Local Sources

BCs for from CMAQ 36/12 km regional modeling

12/4/1 km two-way interactive grid nesting– Allows for recirculation of

pollutants within domains

Plume-in-Grid and PSAT used for local sources– Subgrid-scale sampling grid

use to sample puffs before released to grid

– Ability for near-source “fence line” impacts 300 350 400 450 500 550 600 650 700 750 800

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

Preliminary Results for Total PM2.5

PSAT_PIG - P25

0

5

10

15

20

25

Q1

Q2

Q3

Q4

AN

N

Q1

Q2

Q3

Q4

AN

N

Q1

Q2

Q3

Q4

AN

N Q1

Q2

Q3

Q4

AN

N Q1

Q2

Q3

Q4

ANN Q1

Q2

Q3

Q4

AN

N

Q1

Q2

Q3

Q4

AN

N

Washington Ave. Granite City East St. Louis Ferguson Blair St. Margaretta 2nd & Mound

Con

cent

ratio

n BC

IC

Surface sources

Other PointSources

Source#1

Source#2

Live Puff

Local Sources contributing ~7 and ~3.4 µg/m3 to annual average PM2.5 at Washington Street and Granite City monitoring sites

St. Louis Conceptual ModelTurner and co-workers (2007a,b,c,d) have developed a Conceptual Model for PM2.5exceedances in the St. Louis area– They found that local sources contribute ~3.2

µg/m3 to PM2.5 at the Granite City monitor on average

– The CAMx 12/4/1 km PiG modeling attributes 3.4 µg/m3 to local sources at Granite City

– Turner, J.R. and J.L. Garlock. 2007a. A Conceptual Model for Ambient Fine Particulate Matter over the St. Louis Area. Revision 3.0. Washington University, St. Louis, MO. October.

– Turner, J., J. Garlock, S. Raffuse, J. Rubin, S. Brown, B. Anderson and G. Norris. 2007b. Task#5 Transport Regimes Analysis, Version 1. Technical Memorandum. Washington University, St. Louis, MO. October.

– Turner, J.R., J. Garlock, J. Jaeckels and J. Schauer. 2007c. Task#2 Fine Particulate Matter Carbon Apportionment, Version 1. Washington University, St. Louis, MO. October.

– Turner, J. and J. Garlock. 2007d. Task#4 Urban/Regional Contrast and Intraurban Variability, Version 1. Washington University, St. Louis, MO. October.

Was

hing

ton

Ave.

Gra

nite

City

Eas

t St.

Loui

sFe

rgus

onBl

air S

t.M

arga

retta

2nd

& M

ound

ANN

BC

IC

Surface sources

Other PointSources

Source#1

Source#2

Live Puff0

5

10

15

20

25

Con

cent

ratio

n

ConclusionsRecent advances in PGMs make them more suitable for assessing “single source” contributions to ozone, PM2.5, visibility and deposition– Fine resolution grids, two-way grid nesting, and flexi-nesting– Full chemistry Plume-in-Grid modules– Ozone and PM source apportionment– Full gas-phase and aqueous-phase chemistry and aerosol

thermodynamic modules

The use PGM modeling to assess “single source” air quality, visibility and deposition issues has become more routine: – ASIP point source PM2.5 assessment– Oil and gas AQ and AQRV assessments as part of NEPA– Texas and Arkansas BART assessment– PM2.5 SIP modeling