Modeling Application Process Session 2 Laura Boothe Mike Abraczinskas George Bridgers NC Division of Air Quality Attainment Planning Branch September 30,

Modeling Application Process

Session 2

Laura Boothe Mike AbraczinskasGeorge Bridgers

NC Division of Air QualityAttainment Planning Branch

September 30, 2004

Topics of Discussion

Reasoning behind photochemical grid modeling approach to SIP development

Photochemical grid modeling overview

Illustrate how modeling can be used to evaluate control strategies for various applications

NC’s Early Action Compact (EAC) Ozone modeling results

State of the Air – O3

2001-2003 Ozone Design Values – Each Monitor

2001-2003 Ozone Design Values(Highest value per county)

What does Non-Attainment mean?

EPA label saying air quality does not meet health standards

Requires State to develop a plan showing what control measures are needed to meet the standard

Requires transportation conformity

Requires new source review on new major industrial sources and major modifications on existing sources

Nonattainment Boundaries

Evaluate which monitors are violating Determine appropriate boundaries Conduct public meetings Coordinate with other agencies

impacted by nonattainment designations (NCDOT and NCDOC, Local Government)

State Implementation Plan(SIP)

States are required to develop a strategy or implementation plan to demonstrate attainment of any criteria pollutant violating the NAAQS

NC moved forward with 8-hour ozone SIP planning despite legal issues with the actual 8-hour ozone standard

Recognition that most of NC would not attain the 8-hour ozone standard when designations finally occur

Goals of NC SIP Development

To identify the appropriate emission controls necessary to attain the 8-hour ozone standard statewide

To define any additional emission controls necessary to address the 1-hour ozone violations in Charlotte

To demonstrate a robust attainment strategy that will maintain compliance in the face of future year growth

Why Use a Photochemical Grid Modeling System?

More complex than trajectory style analysis or source dispersion modeling because of detailed chemistry and multiple emission sources (point, mobile, and area)

Ability to establish a base/control case that is verified with known data and the ability to vary the emission inputs of this base to assess control strategies

Detailed process analysis can be performed to follow every chemical process back to originating point(s)

Photochemical Grid Modeling System

Meteorological Model Emissions Processor

Photochemical Model

Temporally and Spatially Gridded Air Quality Output

Data

MM5 SMOKE

MAQSIP

SparseMatrixOperatorKernelEmissions

MultiscaleAirQualitySImulationPlatform

Current Modeling Projects

Early Action Compact (EAC) modeling• Addresses 8-hour ozone• Fayetteville, Triad, Unifour, Mountain (total:

19 counties)

VISTAS modeling• Addresses regional haze• Given the “one-atmosphere” modeling

approach, this will be used for PM2.5 and Ozone to the extent possible

Early Action Compact

On June 19, 2002, EPA Region 6 endorsed Texas’ Protocol for an Early Action Compact

Protocol deals with attainment of the 8-hr National Ambient Air Quality Standards (NAAQS) for ozone

Calls for “early” SIPs and associated controls coupled with a deferral of the effective date of future EPA ozone nonattainment (NA) designations.

Who is Eligible to Participate?

Area may apply for “Early Action 8-hr Compact” if…

• Currently designated attainment of the 1-hour ozone standard

• Air quality monitors show attainment of the 1-hour standard

• Air quality approaches or exceeds the 8-hour standard

Compact Requirements - Milestones and Reporting

Completion of emissions inventories and modeling

Adoption of control strategies that demonstrate attainment

Completion and adoption of the early action SIP revision

Attainment not later than December 31, 2007

Post-attainment demonstration and plan updates

EAC Timeline

December 31, 2002 - Compact signed by all parties in MSA (local officials, state air quality agency, and EPA Region)

June 16, 2003 - Submit list of local control measures being considered

March 31, 2004 - Local plan submitted to the state

April 2004 – EPA designates “Compact Areas” as nonattainment, but defers effective date for these areas

EAC Timeline (continued)

December 31, 2004 - State adopts control measures into SIP & submits to EPA for approval

2005 – Areas implement control measures

June 30, 2006 - Progress assessment and

report to EPA

December 2007 – Areas attain 8-hr ozone

NAAQS

VISTAS

Visibility Improvement State and Tribal Association of the Southeast

Regional Planning Organization established under the 1999 Regional Haze Rule

Collaborative effort of States and Tribes to support management of regional haze and related air quality issues in the Southeastern US.

No independent regulatory authority and no authority to direct or establish State or Tribal law or policy.

VISTAS

• NCDAQ is a technical leader in VISTAS

• Shelia Holman, Technical Analysis Workgroup Chair

• Brock Nicholson, Planning Workgroup Co-Chair

• Mike Abraczinskas, Technical Lead for Meteorological Modeling

• George Bridgers, Technical Lead for Characterizing Meteorology and Conceptual Descriptions

• Laura Boothe, Technical Lead for Emissions


Select areas or domains of interest Select representative ozone episodes Prepare and refine meteorological simulations Prepare and refine emission model inputs Apply photochemical modeling system Performance evaluation on episodes Prepare current and future year emissions

(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test


Select areas or domains of interestSelect areas or domains of interest Select representative ozone episodes Prepare and refine meteorological simulations Prepare and refine emission model inputs Apply photochemical modeling system Performance evaluation on episodes Prepare current and future year emissions


North Carolina 1995 MAQSIP Domain



Grid Structure

Grid Structure, 4km grid spacing


Select areas or domains of interest Select representative ozone episodesSelect representative ozone episodes Prepare and refine meteorological simulations Prepare and refine emission model inputs Apply photochemical modeling system Performance evaluation on episodes Prepare current and future year emissions


Episode SelectionFor EAC Modeling

Draft USEPA guidance suggests:

Variety of meteorological scenarios when 8-hr maxima exceed 84 ppb

Choose episodes containing days with observed 8-hr maxima “close to” (+/- 10 ppb) the design value straddling the period from which the episode was drawn

• Minimum of 3 days

Final guidance may suggest much longer periods.Would apply to non EAC ozone modeling.

Episode SelectionFor EAC Modeling

Four unique ozone episodes selected• July 10-15, 1995• June 20-30, 1996 (Broken into two episodes)• July 10-15, 1997

Regional 1 and 8-hour ozone exceedences during all four episodes

Each episode spread across multiple days Variety of meteorological scenarios

(Stagnate High Pressure, Pre/Post Frontal Passage, Tropical Wx Influence)

8-hour ozone maximums were representative of the design values at the monitoring sites throughout NC

Looking ahead…Episode SelectionFor 8-hour ozone and PM2.5

8-hour ozone modeling analyses will likely include an entire ozone season

• May – September 2002

PM2.5 modeling analyses will be done for the entire year of 2002


Select areas or domains of interest Select representative ozone episodes Prepare and refine meteorological simulationsPrepare and refine meteorological simulations Prepare and refine emission model inputs Apply photochemical modeling system Performance evaluation on episodes Prepare current and future year emissions


Meteorological Modeling

Gridded and hourly variables are needed to simulate advection, diffusion, deposition, chemical transformation, etc.•Wind•Temperature•Water-vapor concentration•Pressure•Vertical diffusivity (effective mixing height)

•Cloud cover•Rainfall rate

Meteorological Modeling

Several iterations may be needed to simulate the meteorological mechanisms that are important to high ozone events

Consider NC’s diverse geographical landscapes• Physical parameterizations that may work well

in the NC Mountains may not work in the NC Coastal Plain

Compromise, compromise, compromise

Meteorological Model

Meteorological Model Performance (Example from EAC modeling)

Meteorological Model Performance (Example from EAC modeling)

Meteorological Model Performance (Example from VISTAS modeling)

Meteorological Model Performance (Example from VISTAS modeling)


Select areas or domains of interest Select representative ozone episodes Prepare and refine meteorological simulations Prepare and refine emission model inputsPrepare and refine emission model inputs Apply photochemical modeling system Performance evaluation on episodes Prepare current and future year emissions


Emissions Requirements

Hourly, gridded, speciated emissions are needed• Point sources:Point sources: utilities, refineries, industrial

sources, etc.

• Area sources:Area sources: gas stations, dry cleaners, fires, etc.

• Motor vehicles:Motor vehicles: cars, trucks, buses, etc.

• Nonroad mobile sources:Nonroad mobile sources: agricultural equipment, recreational marine, lawn mowers, construction equipment, etc.

• Biogenic:Biogenic: trees, vegetation, crops

Emission Processing

Gridding

Speciation

Temporal

Emission Inventory

SMOKE Emission

Model

Photochemical Model

Gridding Allocates emissions to the cells of the grid

• Created for each grid modeled

• Based on 1990 census data (for EAC modeling)

• 2000 census data for VISTAS modeling

Added gridding surrogates for mobile• Travel Demand Areas - created surrogates based

on TDM link data for all 12 road types (for EAC modeling)

• Rest of NC - created surrogates based on NCDOT digitized data for top 6 road types (for EAC modeling)

Speciation

Converts Volatile Organic Compounds (VOC’s) into carbon bond IV species

Updated default profiles to reflect default changes made EPA’s SPECIATE model

Created new speciation profiles for wood furniture finishing based on NC source specific data

Temporal

Adjusts the emissions to the month of the year, day of the week and to the hour of the day

Adjusted highway mobile source profiles to reflect weekly and hourly profiles provided by NCDOT

Old EPA default weekday diurnal profile for Mobile sources

All road types

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23Hour

VM

T %

All road types

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

9.0%

10.0%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Hour

% V

MT

110

130

150

170

190

210

230

250

270

290

310

330

Weekday diurnal profile for MobileBased on data from NCDOT

Used in EAC modeling

Weekend diurnal profile for Mobile

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

9.0%

10.0%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Hour

% V

MT

110

130

150

170

190

210

230

250

270

290

310

330

Used in EAC modeling

Emission Inventory 5 Basic Source Categories

•Point Sources•Area Sources•Nonroad Mobile Sources•Highway Mobile Sources•Biogenics

4 Types of Inventories•Base Case or Episodic •Current Year•Future Year•Control Strategy

Sensitivities can be done on any existing inventory

Point SourcesEpisodic Base Case EAC

North Carolina• Started with ‘95 Ozone Transport Assessment

Group (OTAG) inventory• Adjusted large NOx and VOC non-utility source

emissions to match NCDAQ’s Air Quality Emission Inventory (AQEI)

• Utility source emissions are Continuous Emissions Monitoring (CEM) data for actual episode days

Other States• SIP Call ‘95 base inventory for non-utility• CEM data for most utilities

Area SourcesEpisodic Base Case EAC

North Carolina• Calculated based on NC State specific data

(i.e., employment data, population, etc.)• Backcasted the 2000 EI to episode periods

using BEA data.

Other States• SIP Call ‘95 base inventory

Nonroad Mobile Sources Episodic Base Case EAC

Railroad Locomotives & Aircraft Engines• NC - Based on State specific data (i.e., diesel fuel

consumed, landing/take off data)• Other States - SIP Call ‘95 base inventory

Other Nonroad Equipment• Based on draft version of EPA’s NONROAD 2002

mobile model

Highway Mobile SourcesVehicle Miles Traveled (VMT) Episodic Base Case EAC

North Carolina• NCDOT VMT for areas without Travel

Demand Models (TDM)• TDM VMT for Durham, Forsyth, Guilford,

Mecklenburg, Orange, & Wake Counties and partial Davidson & Randolph Counties

• Neighboring States data supplied by State Air Agencies or State DOT’s

Other States• OTAG ‘95 VMT data

Highway Mobile SourcesMOBILE6 default changes

Area speed studies used adjust morning & afternoon peak speeds

Inspection & Maintenance fractions calculated based on accident data

Correction to vehicle mix to represent increase Sports Utility Vehicles

State/Area specific vehicle age distribution

Biogenic Sources

Estimated using BEIS3 model

Uses gridded land use data

10 meter temperatures


Select areas or domains of interest Select representative ozone episodes Prepare and refine meteorological simulations Prepare and refine emission model inputs Apply photochemical modeling systemApply photochemical modeling system Performance evaluation on episodesPerformance evaluation on episodes Prepare current and future year emissions


Photochemical Model

Mathematically simulates the following processes:

•Emission of chemical precursors (anthropogenic and biogenic)•Advection and diffusion (transport)•Photochemistry•Deposition

Application of MAQSIP

Meteorological and emissions inputs developed

MM5, SMOKE, and MAQSIP models applied for all four base episodes

Rigorous model performance evaluation resulting in multiple reapplications of the models

Model Performance

Model PerformanceExample: 1-hr ozone time series at Enochville (Rowan County)

Model PerformanceExample: 8-hr ozone time series at Enochville (Rowan County)

Model PerformanceExample: 1997 episode Triad monitors

Performance goals (NOT CRITERIA):

Bias = < +/- 20% Norm Gross Error = 30-35%


Select areas or domains of interest Select representative ozone episodes Prepare and refine meteorological simulations Prepare and refine emission model inputs Apply photochemical modeling system Performance evaluation on episodes Prepare current andPrepare current and future year emissionsfuture year emissions

(Projected and Potential Control Strategies)(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test

Current Year Inventories2000 for EAC modeling

Point Sources• Pseudo 2000 for NC• 1999 NEI for other states• Used episodic utility emissions

Area Sources• NC 2000 inventory• Used 1999 NEI for other states

Current Year Inventories2000 for EAC modeling

Nonroad Mobile Sources• NONROAD 2002 model• NC 2000 railroad and airport estimates• Other from NEI 1999

Onroad Mobile Sources• Travel Demand Model (TDM) VMT where

available

• NCDOT “universe” file VMT elsewhere

• Speed data provided by NCDOT.

• Vehicle Age Distribution provided by NCDOT

Future Year Inventories2007 for EAC modeling

Point Sources• Non-utility projected based on USEPA’s EGAS

model growth factors– Some state specific data was used (Furniture

and Textile industries)• NC utility future emissions provided by Duke

Energy and Progress Energy• Other State utility emissions based on USEPA’s

Clear Skies modeling

Area Sources• Projected based on EGAS growth factors

– Some state specific data was used


Nonroad Mobile Sources• EPA’s NONROAD 2002 model• Railroad locomotive emissions projected with

EGAS growth factors• NC’s Aircraft engine emissions projected with

airport specific growth factors (FAA data)


Highway Mobile Sources• In TDM areas, VMT is provided. For other years

that are needed a linear interpolation is performed.

• VMT in other areas is projected using a growth rate from a linear regression completed on the last 10 years of data. The source of this data is NCDOT’s “universe” file.

• Speed data provided by NCDOT.

• Vehicle Age Distribution provided by NCDOT

– The same age distribution is used for the current year and future years… Mobile6 automatically “ages” the fleet.

Future Years Modeled

2007 EAC Attainment Year(control measures listed on next 2 slides)

2012 Maintenance Year 2017 Maintenance Year

Future Control Measures Modeled

Highway Mobile• Tier 2 Standards (2004)• Low Sulfur Gasoline (2006)• Heavy Duty Engine Standard (2007)• Clean Diesel Fuel (2007)• NC’s NOx I/M Program (2002-2006)

Off-Road Mobile• Nonroad Diesel Engine Tier 2 Standards (2001 - 2006)• Nonroad Diesel Engine Tier 3 Standards (2006 - 2008)• Nonroad Spark Engine Phase I Standards (2001-2007)• Heavy Duty Engine Standard (2007)

Future Control Measures Modeled

Point• NOx SIP Call (2004-2006)• NC Clean Smokestacks Act (2007-2013)• RACT, MACT, etc controls

Area• Open Burning Rules (no burning on Ozone Action Days,

2004)• Federal Standards

– Consumer solvents, architectural coatings, etc.

North Carolina 2000 VOC Emissions

Area3%

Point2%

Nonroad Mobile2%

Highway Mobile3%

Biogenics90%

0

200

400

600

800

1000

1200

Point Highway M obile Nonroad M obile Area

ton

s/d

ay

2000 2007 2012 2017

North Carolina NOx EmissionsEAC Modeling

NC Statewide NOx Emissions

Point52%

Highway Mobile36%

Nonroad Mobile11%

Area1%

2000 NOx Emissions

Point35%

Highway Mobile45%

Nonroad Mobile18%

Area2%

2007 NOx Emissions

North Carolina Mobile Emissions EAC Modeling

2000 UrbanMobile NOx Emissions

HDDV57%

HDGV6%

LDGT112%

LDGT25%

LDGV19%

Other1%

2007 Urban Mobile NOx Emissions

HDDV52%

HDGV7%

LDGT117%

LDGT28%

LDGV15%

Other1%

Mobile variables and data

Vehicle Miles Traveled Vehicle age distribution Vehicle type mix Road type Speeds Fuel RVP Temperatures Year being modeled I/M program

• OBDII, tailpipe test– Stringency, Compliance, Waiver Rates

I/M Fractions• Derived from accident data

Temporal profiles Gridding surrogates

Mobile variables and dataExample mobile input file

Mobile QA proceduresINPUTS Check it twice! File parameters are checked by someone

other than the person that developed the file

OUTPUTS Check model logs for errors Comparisons to previous model runs (when

multiple iterations are necessary) Reality checks Spatial consistency/completeness Trends 2017<2012<2007

Mobile QA procedures 2007 NC NOx

NOx tons/day

Wak e Co M e ck le nburg Co

Cook Co

Wayne Co

-50.00

0.00

50.00

100.00

150.00

200.00

250.00

0 200 400 600 800 1000 1200 1400 1600 1800

Mobile QA procedures2007 NC NOx

NOx tons/day

Guilford Co

M eck lenburg Co

Wak e Co

-5.00

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

0 10 20 30 40 50 60 70 80 90 100

Mobile QA procedures 2017 < 2012 < 2007 < 2000



Mobile QA procedures 2012 minus 2007 NC NOx



(Projected and Potential Control Strategies) Re-apply photochemical modeling systemRe-apply photochemical modeling system Analyze the effectiveness of control strategiesAnalyze the effectiveness of control strategies Apply the attainment test

Interpreting Model Results Focus on relative changerelative change between the

current year (2000) in the upper left and the future years (2007, 2012, 2017) in the other 3 panels.

Do notDo not focus on absolute values.

1996 Episode Episode Max

Current Year Results (2000) Attainment Year Results (2007)

5-Year Maintenance Results (2012)


1997 Episode Episode Max

Current Year Results (2000) Attainment Year Results (2007)





(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment testApply the attainment test

Relative vs. Absolute Why use model estimates in a “relative” rather

than “absolute” sense?

• The form of the 8-hr standard (4th highest averaged over 3 years) makes it difficult to tell whether or not a modeled exceedance obtained on one or more days selected from a limited sample of days is consistent with meeting the NAAQS

• Problems with model performance are reduced (although good model performance remains a prerequisite for use of a model in an attainment demonstration)

Attainment Test

DVF = RRF * DVC

DVF = Future Design Value

RRF = Relative Reduction Factor

DVC = Current Design Value

DVC is basedon observeddata

RRF is basedon modeleddata

Future modeled valuesCurrent modeled values

If DVF is 84 ppb, the test is passed.

Attainment Test

Step 1: Compute a current site-specific design value (DVC) from monitored data… If DVC > 75 ppb, then proceed to step 2

Step 2: Use air quality modeling results to estimate a site-specific relative reduction factor (RRF)

Step 3: Multiply the relative reduction factor obtained in step 2 times the site-specific design value in step 1… The result is

a predicted site-specific future design value (DVF)… If DVF is 84 ppb, the test is passed.

DVF = RRF * DVC

EAC Modeling Resultsupdated 9/29/04

Summary of results•By 2007… All EAC areas show attainment

While the EAC modeling can not be used for the Charlotte, Triangle, RMT, GSMNP attainment demonstrations, one can get an idea of the range of ozone concentrations that we might be dealing with in the future years of 2007, 2012, 2017.

EAC Modeling Resultsupdated 9/29/04

Average Ozone Design Value Reduction (ppb)

Hickory Triad Triangle Charlotte Asheville (ridges)

Fayetteville Down East Asheville (valleys)

Others

200714141312121010109

2012 19

1819181715141412

2017212122211919161615

NC’s Planning Schedule

Aug 2004 – Begin PM2.5 and 8-hr ozone base year modeling through VISTAS

Nov 2004 – Begin modeling PM2.5 and 8-hr ozone future years

Dec 2004 – Submit SIP to USEPA for EAC areas All of 2005 – Complete PM2.5 and 8-hr ozone modeling

with control strategies Jan 2006 – Start rule making process for CLT area (RFP

15% VOC requirements) Jun 2007 – Submit SIP to USEPA for 8-hr ozone Jan/Feb 2008 – SIP due to USEPA for PM2.5

nonattainment areas (NCDAQ will likely submit this with 8-hr ozone SIP)

For More Information:

Visit our web site:• http://ncair.org

Technical web site:• http://www.cep.unc.edu/empd/projects2/NCDAQ/PGM/results/

VISTAS information:• http://www.vistas-sesarm.org/

Modeling Application Process Session 2 Laura Boothe Mike Abraczinskas George Bridgers NC Division of Air Quality Attainment Planning Branch September 30,

Documents

state n

charlotte n

approval n

control measures n

epa region n

n epa label

naaqs n nc

eac timeline n