EPA/FLM Demonstration Project€¦ · EPA/FLM Single Source LRT Demonstration Project Bret A. Anderson USDA Forest Service 10th Conference on Air Quality Modeling EPA‐Research Triangle

EPA/FLM Single Source LRT Demonstration Project

Bret A. AndersonUSDA Forest Service

10th Conference on Air Quality Modeling10 Conference on Air Quality ModelingEPA‐Research Triangle Park, NC Campus

March 13‐15, 2012

OutlineOutline

• Discuss use of models for AQ/AQRV underDiscuss use of models for AQ/AQRV under NEPA

• Discuss design elements of the EPA/FLM singleDiscuss design elements of the EPA/FLM single source model evaluation project

• Examine initial results of the evaluationExamine initial results of the evaluation project

• Discuss practical considerations of use ofDiscuss practical considerations of use of Eulerian models more routinely in a regulatory framework

NEPA Requirements and the FLM’sNEPA Requirements and the FLM s

• Air quality modeling for NEPA was well definedqua ty ode g o as e de ed– AERMOD for near‐field analyses– CALPUFF for far‐field

• …And then our monitors started to find problems– Winter ozone in Upper Green River Basin in Wyoming

d hand Uinta Basin in Utah

• We now had to address ozone air quality modeling for pollutants that occurred in times ofmodeling for pollutants that occurred in times of the year and in remote locations that were once considered only urban, summertime issues. y ,

The Experience of the Multi‐Use Land Management Agencies

• Multiple use land management agencies such as the Bureau of Land Management (BLM) and US Forest Service (USFS) are responsible for developing environmental impact statements (EIS’) for any resource management d i i th t d f f d l l ddecisions that are made for federal lands.

• For air quality, this means we must analyze potential impacts to local both local and regional air quality for each

t ti th t i id d Thiresource management option that is considered. This translated into running:– AERMOD for near‐field NAAQS/increment analysis

CALPUFF f f fi ld NAAQS/i d AQRV– CALPUFF for far field NAAQS/increment and AQRV– CAMx/CMAQ for ozone NAAQS

The FLM Experience (continued)• The complexity and the cost associated with meeting the needs of air quality

analyses under NEPA have grown considerably in the last five years.

• In response USDA DOI and USEPA entered into a memorandum of understandingIn response, USDA, DOI, and USEPA entered into a memorandum of understanding outlining generally agreed upon procedures for conducting air quality analyses. Principals of the MOU are to:

– Establish agreed upon procedures for conducting air quality analyses, formal stakeholder input process, and dispute resolution procedures

– Reduce the costs to both agencies and development project proponents through promotion of modeling techniques which allow for leveraging i ti l i t th t t ti lexisting analysis to the extent practical

• Reduction of the burden in modeling can occur through two approaches:

1. Establishing a reusable modeling framework – regional air quality analyses that bracket development potential in a given airshed that can be leveragedthat bracket development potential in a given airshed that can be leveraged to describe potential impacts for an individual project

2. Promoting use of single modeling platforms to the extent practical to deal with ozone and PM2 5 NAAQS and AQRV requirements of NEPA. 2.5 q

Where Does EPA and FLM’s Go From Here?

• NEPA requirements and Sierra Club petition equ e e ts a d S e a C ub pet t onecessitate that both FLM’s and EPA reassess the suitability of the existing modeling paradigm to

• In order to address these needs, the EPA and FLM’s undertook a project to compare the model

di ti f i ti d l d ipredictions of existing models and emerging models to understand both the predicted impacts for resource management decisions and to betterfor resource management decisions and to better understand the resource requirements and challenges to implementation.

EPA WA Task 6: lSingle‐Source LRT Demonstration

• Apply LRT chemical dispersion models for example test sources as one would for a PSD far field Class I assessmentone would for a PSD far‐field Class I assessment– 2005 and 2006 annual simulations examined– CALPUFF/CALMET and CALPUFF/MMIF– CAMx using PSAT/APCA source apportionment– CAMx using PSAT/APCA source apportionment

• Compare far‐field air quality and air quality related values (AQRVs) metrics at Class I areas across LRT dispersion models– Maximum concentrations SO2 NO2 PM10)Maximum concentrations SO2, NO2, PM10)– Maximum visibility impairment (FLAG, 2010)– Maximum sulfur and nitrogen deposition

• This presentation documents LRT dispersion model simulations and consequence analyses performed by ENVIRON under contract to EPA in 2011ENVIRON under contract to EPA in 2011.

7

Task 6: Single‐Source LRT Model Demonstration

• 2005 12/4 km FCAQTF • 2006 12 km UT‐CO720 2005 12/4 km FCAQTF 2006 12 km UT CO

240

360

480

600

720

360

-240

-120

0

120

-840

-720

-600

-480

-360

12 km

4 km

-2280-2160-2040-1920-1800-1680-1560-1440-1320-1200-1080 -960 -840 -720 -600 -480 -360

New Mexico CAMx 12 km Nested Domain

IMPROVECASTNET

12 km

8

CASTNETNADP

CAMx 12 km: 167 x 137* (-2316, -912) to ( -312, 732)CAMx 04 km: 101 x 92* (-1192, -508) to ( -788, -140)

* includes buffer cells

Test Sources

2005 FCAQTF 12/4 km

• 5 EGU Point Sources

• 9 oil and gas point and g parea source areas– 9 x 9 4 km area sources

Any point sources in– Any point sources in region

9

T t STest Sources2006 12 km UT‐CO

• 13 EGU Point Sources

• 11 oil and gas point and area source areas– 3 x 3 12 km grid cells– Any point sources in regionregion

10

2005 4 km Annual NO2 (top) and SO2 (bottom) at Class I receptors: FCAQTF(bottom) at Class I receptors: FCAQTFCALPUFF/CALMET vs. CAMx CALPUFF/CALMET vs. CALPUFF/MMIF CAMx vs. CALPUFF/MMIF

y = 1.0469x - 0.00240.6

Annual NO2 (ug/m3)

y = 0.6521x - 0.0020.6

Annual NO2 (ug/m3)

y = 1.5724x + 0.00160.6

Annual NO2 (ug/m3)

0.2

0.3

0.4

0.5

CAM

x N

O2

(ug/

m3)

0.2

0.3

0.4

0.5

PU

FF/M

MIF

NO

2 (u

g/m

3)

0.2

0.3

0.4

0.5

CAM

x N

O2

(ug/

m3)

0

0.1

0 0.1 0.2 0.3 0.4 0.5 0.6CALPUFF/CALMET NO2 (ug/m3)

0

0.1

0.0 0.1 0.2 0.3 0.4 0.5 0.6

CAL

CALPUFF/CALMET NO2 (ug/m3)

0

0.1

0 0.1 0.2 0.3 0.4 0.5 0.6CALPUFF/MMIF NO2 (ug/m3)

y = 1.0708x - 0.0004

0.4

0.5

0.6

m3)

Annual SO2 (ug/m3)

y = 0.7204x - 0.0043

0.25

0.3

0.35

0.4

(ug/

m3)

Annual SO2 (ug/m3)

y = 1.4737x + 0.0064

0.4

0.5

0.6

m3)

Annual SO2 (ug/m3)

0.1

0.2

0.3

CAM

x SO

2 (u

g/m

0.05

0.1

0.15

0.2

CALP

UFF/

MM

IF S

O2

0.1

0.2

0.3

CAM

x SO

2 (u

g/m

11

00 0.1 0.2 0.3 0.4 0.5 0.6

CALPUFF/CALMET SO2 (ug/m3)

00 0.1 0.2 0.3 0.4


00 0.1 0.2 0.3 0.4 0.5 0.6

CALPUFF/MMIF SO2 (ug/m3)

2005 4 km 24‐hr (top) and 3‐hr (bottom) SO2 at Class I receptors: FCAQTFSO2 at Class I receptors: FCAQTF

CALPUFF/CALMET vs. CAMx CALPUFF/CALMET vs. CALPUFF/MMIF CAMx vs. CALPUFF/MMIF

1 2049 0 0602

14

High 1st High 24-hour SO2 (ug/m3)

14


y 1 8391x + 0 1091

14


y = 1.2049x - 0.0602

6

8

10

12

CAM

x SO

2 (u

g/m

3)

y = 0.6557x - 0.0925

6

8

10

12

FF/M

MIF

SO

2 (u

g/m

3)

y = 1.8391x + 0.1091

6

8

10

12

CAM

x SO

2 (u

g/m

3)

0

2

4

0 2 4 6 8 10 12 14CALPUFF/CALMET SO2 (ug/m3)

0

2

4

0 2 4 6 8 10 12 14

CALP

UF


0

2

4

0 2 4 6 8 10 12 14CALPUFF/MMIF SO2 (ug/m3)

y = 0.5839x + 0.083

30

35

40

45

3)


y = 0.6641x - 0.4691

30

35

40

45

g/m

3)


y = 0.8636x + 0.53

30

35

40

45

3)


5

10

15

20

25

CAM

x SO

2 (u

g/m

3

5

10

15

20

25

CALP

UFF/

MM

IF S

O2

(ug

5

10

15

20

25

CAM

x SO

2 (u

g/m

3

12

0

5

0 5 10 15 20 25 30 35 40 45CALPUFF/CALMET SO2 (ug/m3)

0

5

0 5 10 15 20 25 30 35 40 45CALPUFF/CALMET SO2 (ug/m3)

0

5

0 5 10 15 20 25 30 35 40 45CALPUFF/MMIF SO2 (ug/m3)

2005 4 km ann (top) and 24‐hr (bottom) PM10 at Class I receptors: FCAQTFPM10 at Class I receptors: FCAQTF


y = 0.2624x + 0.0006

0 25

0.3

Annual PM10 (ug/m3)

y = 0.6922x - 0.0009

0 25

0.3

Annual PM10 (ug/m3)

y = 0.374x + 0.001

0 25

0.3

Annual PM10 (ug/m3)

0.1

0.15

0.2

0.25

CAM

x PM

10 (

ug/m

3)

0.1

0.15

0.2

0.25

ALP

UFF/

MM

IF P

M10

(ug/

m3)

0.1

0.15

0.2

0.25

CAM

x PM

10 (

ug/m

3)

0

0.05

0 0.05 0.1 0.15 0.2 0.25 0.3CALPUFF/CALMET PM10 (ug/m3)

0

0.05

0 0.05 0.1 0.15 0.2 0.25 0.3

CA

CALPUFF/CALMET PM10 (ug/m3)

0

0.05

0 0.05 0.1 0.15 0.2 0.25 0.3CALPUFF/MMIF PM10 (ug/m3)

High 1st High 24-hour PM10 (ug/m3) High 1st High 24-hour PM10 (ug/m3) High 1st High 24-hour PM10 (ug/m3)

y = 0.1745x + 0.0389

3

3.5

4

4.5

5

0 (u

g/m

3)

y = 0.6266x - 0.0111

3

3.5

4

4.5

5

PM10

(ug/

m3)

y = 0.2646x + 0.0466

3

3.5

4

4.5

5

0 (u

g/m

3)

0

0.5

1

1.5

2

2.5

0 0 5 1 1 5 2 2 5 3 3 5 4 4 5 5

CA

Mx

PM

10

0

0.5

1

1.5

2

2.5

0 0 5 1 1 5 2 2 5 3 3 5 4 4 5 5

CA

LPU

FF/M

MIF

P

0

0.5

1

1.5

2

2.5

0 0 5 1 1 5 2 2 5 3 3 5 4 4 5 5

CA

Mx

PM

10

13

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5CALPUFF/CALMET PM10 (ug/m3)

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5CALPUFF/CALMET PM10 (ug/m3)

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5CALPUFF/MMIF PM10 (ug/m3)

Different PM10 species mappings used w/ CALPUFF vs. CAMx/PSAT

2006 12 km annual NO2 (top) and SO2 (bottom) at Class I receptors: UT‐CO(bottom) at Class I receptors: UT COCALPUFF/CALMET vs. CAMx CALPUFF/CALMET vs. CALPUFF/MMIF CAMx vs. CALPUFF/MMIF

y = 0.0672x + 0.00912.8

3

3.2

Annual NO2 (ug/m3)

y = 0.2962x + 0.00642.8

3

3.2

Annual NO2 (ug/m3)

y = 0.4987x + 0.00462.8

3

3.2

Annual NO2 (ug/m3)

0 8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

CAM

x NO

2 (u

g/m

3)

0 8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

CALP

UFF/

MM

IF N

O2

(ug/

m3)

0 8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

CA

Mx

NO

2 (u

g/m

3)

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2


0

0.2

0.4

0.6

0.8

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2

C


0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2

CALPUFF/MMIF NO2 (ug/m3)

Annual SO2 (ug/m3) Annual SO2 (ug/m3) Annual SO2 (ug/m3)

y = 0.389x + 0.003

0.2

0.25

0.3

0.35

0.4

2 (u

g/m

3)

y = 0.4628x + 0.0021

0.2

0.25

0.3

0.35

0.4

IF S

O2

(ug/

m3)

y = 0.9331x + 0.0008

0.2

0.25

0.3

0.35

0.4

2 (u

g/m

3)

0

0.05

0.1

0.15

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

CAM

x S

O2


0

0.05

0.1

0.15

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

CAL

PUF

F/M

MI

0

0.05

0.1

0.15

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

CAM

x S

O2

14

CALPUFF/CALMET SO2 (ug/m3) CALPUFF/CALMET SO2 (ug/m3) CALPUFF/MMIF SO2 (ug/m3)

Very high CALPUFF/CALMET (CALPUFF/MMIF) outlier

2006 12 km 24‐hr (top) and 3‐hr (bottom) SO2 at Class I receptors: UT‐CO(bottom) SO2 at Class I receptors: UT CO


4


4


4


y = 0.3689x + 0.0205

1 5

2

2.5

3

3.5

AM

x SO

2 (u

g/m

3)

y = 0.5301x + 0.0224

1 5

2

2.5

3

3.5

FF/M

MIF

SO

2 (u

g/m

3)

y = 0.6652x + 0.0077

1 5

2

2.5

3

3.5

AM

x SO

2 (u

g/m

3)

0

0.5

1

1.5

0 0.5 1 1.5 2 2.5 3 3.5 4

CA


0

0.5

1

1.5

0 0.5 1 1.5 2 2.5 3 3.5 4

CAL

PUF


0

0.5

1

1.5

0 0.5 1 1.5 2 2.5 3 3.5 4

CA


y = 0.2432x + 0.046

10

12

14

)


y = 0.4807x + 0.0735

10

12

14

ug/m

3)


y = 0.5059x + 0.0088

10

12

14

)


2

4

6

8

CAM

x SO

2 (u

g/m

3)

2

4

6

8

CALP

UFF/

MM

IF S

O2

(u

2

4

6

8

CAM

x SO

2 (u

g/m

3)

15

00 2 4 6 8 10 12 14


00 2 4 6 8 10 12 14


00 2 4 6 8 10 12 14


2006 12 km CALPUFF High “Outlier”2006 12 km CALPUFF High Outlier

• Maximum annual NO2 and SO2 by CALPUFF/CALMET a u a ua O a d SO by U /occurs for smallest EGU12 (13 TPY NOX and 1 TPY SO2)

• EGU12 located within Holy Cross Wilderness (Class II area) so likely nearly co‐located with receptors

• Maximum CALMET, MMIF and CAMx annual NO2 t ticoncentrations:

– 3.1; 0.6; and 0.02 µg/m3

• We understand why CALMET much higher than CAMx• We understand why CALMET much higher than CAMx– CAMx configured for LRT application with 12 km grid

• Why MMIF and CALMET so different?y

16

CALMET (blue) vs. MM5 (red) Surface Wind Fi ld 2006 12k UT COFields: 2006 12km UT‐CO

17

CALMET (blue) vs. MM5 (red) 480 m Wind Fields: 2006 12km UT COFields: 2006 12km UT‐CO

18

EGU12 Holy Cross Wilderness OutlierEGU12 Holy Cross Wilderness Outlier

• CALMET modifying and slowing MM5 windsy g g• Occurs at the surface and aloft• Unclear whether diagnostic effects or

( )observation Objective Analysis (OA) procedure is doing this

• Occurs throughout the year• Occurs throughout the year• Results in very high concentrations in CALPUFF/CALMET/

• Better agreement with CAMx and CALPUFF/CALMET at other sites

19

2005 (top) and 2006 (bottom) VisibilityCALPUFF/CALMET vs. CAMx CALPUFF/CALMET vs. CALPUFF/MMIF CALPUFF/MMIF vs. CAMx

y = 0.2301x + 0.3492

40

45

50

High 1st High Extinction from Sources (Mm-1)

y = 0.6296x - 0.0007

40

45

50


y = 0.3452x + 0.3991

40

45

50


15

20

25

30

35

CAM

x B

ext

(Mm

-1)

15

20

25

30

35

PUFF

/MM

IF B

ext (

Mm

-1)

15

20

25

30

35

CAM

x B

ext

(Mm

-1)

0

5

10

15

0 5 10 15 20 25 30 35 40 45 50CALPUFF/CALMET Bext (Mm-1)

0

5

10

15

0 5 10 15 20 25 30 35 40 45 50

CALP

CALPUFF/CALMET Bext (Mm-1)

0

5

10

15

0 5 10 15 20 25 30 35 40 45 50CALPUFF/MMIF Bext (Mm-1)

CALPUFF/CALMET vs. CAMx CALPUFF/CALMET vs. CALPUFF/MMIF CALPUFF/MMIF vs. CAMx

y = 0.6632x + 0.093940

45


y = 0.7572x + 0.258840

45


y = 0.6034x + 0.17740

45


15

20

25

30

35

CAM

x Be

xt (

Mm

-1)

15

20

25

30

35

PUFF

/MM

IF B

ext (

Mm

-1)

15

20

25

30

35

CAM

x Be

xt (

Mm

-1)

200

5

10

0 5 10 15 20 25 30 35 40 45CALPUFF/CALMET Bext (Mm-1)

0

5

10

0 5 10 15 20 25 30 35 40 45

CAL

CALPUFF/CALMET Bext (Mm-1)

0

5

10

0 5 10 15 20 25 30 35 40 45CALPUFF/MMIF Bext (Mm-1)

2005 4 km Maximum Visibilityy• Spatial variability across receptors (CALPUFF) and grid cells that intercept Class I area– Except for close source‐receptors, variability similar

21

Nitrogen Deposition (kg‐N/ha/yr)

• CAMx N deposition much• CAMx N deposition much greater (factor of ~2) than CALPUFF

• CALPUFF/MMIF N deposition slightly greater than CALPUFF/CALMETCALPUFF/CALMET

22

Sulfur Deposition (kg‐S/ha/yr)

• Sulfur deposition comparable– CAMx > CALPUFF/CALMET

– CPUFF/MMIF > CALPUFF/CALMET– CPUFF/MMIF > CALPUFF/CALMET

– Except for one point: CAMx=CALPUFF/MMIF

( i il lt f 2006)• (similar results for 2006)

23

Why CAMx Estimating Higher N Deposition than CALPUFF?

• Different species mappings with CAMx more N species and only including NH4 from test source and CALPUFF including NH4 assuming SO4/NO3 neutralizedassuming SO4/NO3 neutralized

• Performed CAMx species mapping using CALPUFF rules– Wrong direction for large sources

0.3000.4000.5000.600

‐N/ha]

Annual Nitrogen Deposition (Wet + Dry)EGU1

camx

camx puffN

calpuff

– Not enough small sources

0.0000.1000.200

ande

lier N

M

La Garita

Wild

erne

ss

sa Verde

NP

rifie

d Forest

NP

Pedro Parks

Wild

Wem

inuche

Wild

erne

ss

[kg‐

calpuff mmif

Annual Nitrogen Deposition (Wet + Dry)EGU3

Ba

Mes

Petr

San P W W

Class I Area

0.0000.0050.0100.0150.0200.0250.0300.035

M a ss NP

NP ks

he

ss

[kg‐N/ha]

EGU3

camx

camx puffN

calpuff

calpuff mmif

Band

elier N

M

La Garita

Wild

erne

Mesa Ve

rde N

Petrified

Forest N

San Pedro Park

Wild

Wem

inuch

Wild

erne

s

Class I Area24

Why CAMx estimating higher N Deposition than CALPUFF?Why CAMx estimating higher N Deposition than CALPUFF?

• CAMx carrying more N as nitric acid (HNO3)y g ( )• CALPUFF carrying more N as particle nitrate (NO3)

• CALPUFF default ammonia background 10 ppb• CAMx simulates ammonia/ammonium

Typically estimates values << 1 ppb in high terrain of– Typically estimates values << 1 ppb in high terrain of Rocky Mountains

• More representative of measurements (e.g., Dinosaur National Monument)National Monument)

• Revised CALPUFF runs with more realistic background ammonia

25

Barriers to Implementation ‐lComputational

• The modeling platforms for the permit modelingThe modeling platforms for the permit modeling community are largely Microsoft Windows based and are engineered for serial applications of models.

• The meteorological and photochemical modeling community is largely Unix/Linux based.

• Time necessary for annual PGM simulations compared to current model recommended for AQRV

lanalyses.

Computational ConsiderationsComputational Considerations

• Adaptation of the PGM platforms to operate in a Windows p p pbased environment?– The permit modeling community typically does not have the same

level of fluency in either the Unix/Linux operating system or Fortranlevel of fluency in either the Unix/Linux operating system or Fortran based programming skills that are essential skill requisites

– IT authorities within State and Local permitting agencies often lack the familiarity and/or resources to dedicate to systems administration andfamiliarity and/or resources to dedicate to systems administration and security for Unix/Linux based systems, and thus actively prevent the acquisition of such equipment or, if such hardware is acquired, prevent the presence of such equipment on the State’s internal p p q pnetwork.

Barriers to Implementation ‐lRegulatory

• The operational construct for the permit modeling community is highly rigidcommunity is highly rigid– Based upon a series of regulations and guidelines which restrict operational flexibility in order to promote more general consistency in the application of modelsgeneral consistency in the application of models.

• The operational construct of the meteorological and photochemical modeling communities is vastly diffdifferent– Based upon a more loosely binding set of EPA recommendations which typically encourage adapting b h d d l h d hboth science and modeling techniques to produce the most scientifically feasible answer given the constraints of the state‐of‐the‐science.

Regulatory ConsiderationsRegulatory Considerations

• The differences in the operational paradigms between the two communities will require both the EPA and the FLM’s to develop a more rigid set of operational procedures similar to the current permit modeling paradigm in order to insure both a scientifically sound and consistent set of procedures to prevent an ‘anything goes’ process as would likely develop without such procedures.p

• Length of meteorological record for PGM’s will likely have to be expanded to be consistent with requirements of GAQM (e g 3 years of prognostic data)(e.g. 3 years of prognostic data).

• Development of significance thresholds for single source (cause or contribute test) required for NAAQS demonstrationsdemonstrations.

ConclusionConclusion

• PGM’s capable of assessing single source impacts for both p g g pAQRV and ozone requirements under PSD.

• Source apportionment eliminates need for multiple “zero‐t”out” runs

• Significant barriers remain to implementation of PGM’s

– Increased computational requirementsIncreased computational requirements

– Increased training requirements for permit modeling staff

– Creation of a hybrid regulatory and guidance framework y g y gfor implementation of PGM’s within a regulatory permit modeling paradigm which is highly rigid and prescriptive

Documents

EPA/FLM Demonstration Project€¦ · EPA/FLM Single Source LRT Demonstration Project Bret A. Anderson USDA Forest Service 10th Conference on Air Quality Modeling EPA‐Research Triangle