Modeling Application Process Session 2 Laura Boothe Mike Abraczinskas George Bridgers NC Division of Air Quality Attainment Planning Branch September 30, 2004
Mar 27, 2015
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
Modeling Application Process
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
Modeling Application Process
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
(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test
North Carolina 1995 MAQSIP Domain
North Carolina 1996 MAQSIP Domain
North Carolina 1997 MAQSIP Domain
Grid Structure
Grid Structure, 4km grid spacing
Modeling Application Process
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
(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test
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
Modeling Application Process
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
(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test
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)
Modeling Application Process
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
(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test
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
Modeling Application Process
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
(Projected and Potential Control Strategies) Re-apply photochemical modeling system Analyze the effectiveness of control strategies Apply the attainment test
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%
Modeling Application Process
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
Future Year Inventories2007 for EAC modeling
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)
Future Year Inventories2007 for EAC modeling
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 2007 NC NOx
Mobile QA procedures 2007 NC NOx
Mobile QA procedures 2012 minus 2007 NC NOx
Modeling Application Process
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 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)
10-Year Maintenance Results (2017)
1997 Episode Episode Max
Current Year Results (2000) Attainment Year Results (2007)
5-Year Maintenance Results (2012)
10-Year Maintenance Results (2017)
Modeling Application Process
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 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/