Presentation is loading. Please wait.

Presentation is loading. Please wait.

Air Quality Modeling for VISTAS James W. Boylan, Ph.D. Georgia Department of Natural Resources (VISTAS Technical Lead for Air Quality Modeling) 2005 EPA.

Similar presentations


Presentation on theme: "Air Quality Modeling for VISTAS James W. Boylan, Ph.D. Georgia Department of Natural Resources (VISTAS Technical Lead for Air Quality Modeling) 2005 EPA."— Presentation transcript:

1 Air Quality Modeling for VISTAS James W. Boylan, Ph.D. Georgia Department of Natural Resources (VISTAS Technical Lead for Air Quality Modeling) 2005 EPA Region 4 Modelers Workshop Atlanta, GA March 10, 2005

2 Outline Background VISTAS Phase I Modeling –CMAQ Sensitivity Tests –Emissions Sensitivity Analysis VISTAS Phase II Modeling –Annual Simulations PM 2.5 and Ozone Modeling VISTAS Web Links

3 Background

4 RPOs: created by EPA to initiate and coordinate activities associated with the management of regional haze at federally mandated Class I areas.

5 VISTAS Organization John Hornback (SESARM) – Executive Director Coordinating Committee –VISTAS State Air Directors Workgroups –Data, Planning, and Technical Analysis Workgroups Workgroup Participants –VISTAS State Governments AL, GA, MS, FL, NC, SC, TN, KY, VA, WV –Tribal Governments Eastern Band of Cherokee Indians –Federal Agencies EPA and FLMs –Industry

6 Regional haze is the impairment of visibility caused by the presence of particulate matter in the atmosphere that scatter and absorb light Visibility is a measure of the clearness of the atmosphere –Light Extinction ( b ext ) b ext (Mm -1 ) = 3* f(RH) *[SO 4 ] + 3* f(RH) *[NO 3 ] + 4*[ORG] + 10*[EC] + 1*[Soils] + 0.6*[PMC] + b rayleigh b rayleigh = 10 Mm -1 –Deciview dV=10*ln(b ext /b rayleigh ) Regional Haze

7 Regional Haze Rule Objectives of Regional Haze Rule –Achieve natural (no man-made impairment) visibility conditions at federal mandated Class I areas by 2064 for worst 20% visibility days –No worsening in visibility at Class I areas for best 20% visibility days First progress SIP due April 5, 2008 demonstrating progress toward natural conditions between 2000-2004 and 2018

8 Evaluation of Reasonable Progress Reasonable Progress must be demonstrated every 10 years Natural Background 20% Haziest Days 2000 2018 Year 2064 dV 20% Cleanest Days

9 http://vista.cira.colostate.edu/views/ b ext on 20% Haziest Days (2002)

10 VISTAS Class I Areas

11 b ext on 20% Haziest Days (2000-2002) Mountain Sites Coastal Sites

12 VISTAS Modeling Approach Modeling Systems used by VISTAS –MM5 for meteorological modeling –SMOKE for emissions modeling –CMAQ for air quality modeling Phase I Modeling –Evaluate models for 3 episodes to identify optimal model configuration for annual modeling –Preliminary evaluation of emission sensitivities to help develop control strategies Phase II Modeling –Perform annual modeling for 2002 and 2018 for use in Regional Haze SIPs

13 Future year emissions (e.g., 2018) Compare to Air Quality Goals Emissions control strategy Modeling Complete Pollutant distributions and sensitivities Future year (e.g., 2018) emissions with controls NO YES Reasonable progress and future year modeling Air Quality Model Both modeling runs use the same meteorological & air quality inputs Note: Air Quality Model Pollutant distribution Model Performance Evaluation Base year emissions (e.g., 2002) Base case modeling

14 MM5 Meteorological Observations 3-d model predictionsLand use, surface elevation, etc 3-D Meteorological Fields (temperature, wind speed, wind direction, humidity, etc) CMAQSMOKE Initial and boundary conditions Photolysis rates NRM MOBILE6 TP+ Measured EI EGAS 3-D Pollutant Distributions and 3-D Sensitivities VISTAS Modeling System

15 CMAQ is a Grid-Based Model SiSi SiSi RiRi uiui uiui uiui KiKi KiKi KiKi

16 Modeled Mobile NO Emissions

17 Modeled Wind Vectors

18 Modeled PM 2.5

19 Phase I CMAQ Sensitivities

20 Phase I Modeling Overview Literature Review Emissions Modeling for 3 episodes – SMOKE Air Quality Modeling for 3 episodes – CMAQ –Perform Model Configuration Sensitivity Tests – Recommend Optimal Model Configuration Protocol for Phase II Modeling Technical Web Site –http://pah.cert.ucr.edu/vistas/

21 Phase I Modeling Details Air Quality Modeling Team –Environ International Corporation –University of California – Riverside –Alpine Geophysics, LLC Modeling Episodes –January 1 ‑ 20, 2002 (20 episode days + ramp ‑ up days) –July 13 ‑ 21, 1999 (9 episode days + ramp ‑ up days) –July 13 ‑ 27, 2001 (15 episode days + ramp ‑ up days) Modeling Domain –36 km grid resolution (149 x 113) –12 km grid resolution (169 x 178) –19 vertical layers (collapsed from 34 MM5 layers)

22 VISTAS 36 km Grid

23 VISTAS 12 km Grid

24 Model Performance Evaluation Evaluate for each major component of PM –Sulfate (SO 4 )Nitrate (NO 3 ) –Elemental Carbon (EC)Organic Carbon (OC) –Soil (Other PM 2.5 )Coarse Mass (CM) Evaluate separately across each network –IMPROVE (24-hr speciated PM and PM mass) –CASTNet (Weekly speciated PM, some gas) –STN (24-hr speciated PM) –SEARCH (24-hr and hourly speciated PM/gas) –AQS (Hourly gaseous species: O3, NO2, SO2, CO) –NADP (Weekly wet deposition: SO4, NO3, NH4)

25 Monitoring Networks

26 Summary of Model Performance January 2002 Episode –Sulfate, Elemental Carbon, Organic Carbon, and Coarse Mass in the “Ball Park” –Large Nitrate Overestimation Ammonia Emissions (Magnitude and Temporal Distribution)? Dry Deposition? Chemistry? Nighttime Mixing? Others? –Large Soil Overestimation Emissions (Magnitude and Speciation)? Mixing (PBL Heights)? Others? July 1999 and July 2001 Episodes –Sulfate, Elemental Carbon, Organic Carbon, and Coarse Mass in the “Ball Park” –Nitrate Underestimation –Soil Overestimation

27 1)Fugitive Dust Transport Factor FDTF=1.0 vs. FDTF=0.25 vs. FDTF=0.05 2)Number of Vertical Layers NLAYS=34 vs. NLAYS=19 3)Vertical Diffusivity - Minimum Kz Kz_min=1.0 vs. Kz_min=0.1 4)Ammonia Emissions (Winter Episode) 0% Reduction vs. 50% Reduction Standard Diurnal Pattern vs. Revised Diurnal Pattern 5)Mexican/Canadian Emissions MX/CAN Emissions vs. No MX/CAN Emissions 6)Boundary Conditions EPA Default vs. GEOS-CHEM 7)Boundary Layer Heights – Minimum PBLs Standard PBL Code vs. Revised PBL Code CMAQ Sensitivity Tests

28 8.Alternative MM5 Configuration Pleim-Xiu vs. NOAH-ETA-MY 9.Aerosol Mass Conservation No Patch vs. GT Patch 10.Alternative Chemical Mechanisms CB-IV vs CB4-2002 vs. SAPRC-99 11.Alternative Aerosol Module AE3/ISORROPIA vs. CMAQ–AIM 12.Grid Resolution 36 km vs. 12 km 13. Alternative Air Quality Model CMAQ vs. CAMx CMAQ Sensitivity Tests (cont.)

29 Phase I Emission Sensitivities

30 Approach Georgia Tech performed emission sensitivities using CMAQ on the VISTAS 12 km modeling domain Model simulations for two episodes –July 13-27, 2001 and January 1-20, 2002 –2018 OTB and 2018 OTW (next slide) Brute-force sensitivities performed by reducing specific emissions by 30% Modeling results used in a relative fashion rather than absolute fashion Goal is to perform a PRELIMINARY evaluation of the reasonable progress goals at each Class I area and evaluate the relative importance of various emission reductions

31 Emission Projection Scenarios On-the-Books (OTB) - Promulgated as of July 1, 2004 –Atlanta / Northern Kentucky / Birmingham 1-hr SIPs –Combustion Turbine MACT –Gulf Power SCR application –Heavy Duty Diesel (2007) Engine Standard –Industrial Boiler/Process Heater/RICE MACT –Large Spark Ignition and Recreational Vehicle Rule –Nonroad Diesel Rule –North Carolina Clean Smokestacks Act –NOx RACT in 1-hr NAA SIPs –NOx SIP Call (Phase I) –Petroleum Refinery Initiative –RFP 3% Plans where in place for one hour plans –TECO & VEPCO Consent Agreements –Tier 2 Tailpipe –Title IV for Phase I and II EGUs –VOC 2-, 4-, 7-, and 10-year MACT Standards On-the-Way (OTW) – OTB Assumptions plus: – Clean Air Interstate Rule (CAIR) –NOx SIP Call (Phase II) –8-hr attainment plans (e.g., NOx RACT)

32 MACA Reasonable Progress Goal 30.3 dV = 206.97 Mm -1 25.44 dV = 127.28 Mm -1

33 MACA Required Reductions b ext 2002 – b ext 2018 = 206.97 Mm -1 – 127.28 Mm -1 = 79.69 Mm -1 On the Books Regulations reduces extinction by 18.47 Mm -1 Need an additional reduction of: 79.69 Mm -1 - 18.47 Mm -1 = 61.23 Mm -1

34 Level 1 Sensitivity Acronyms OTB-TYP  2018 OTB – 2002 Typical OTW-OTB  2018 OTW – 2018 OTB ASO2  30% reduction in all SO 2 domain-wide ANOX  30% reduction in all NO X domain-wide ANH3  30% reduction in all NH 3 domain-wide ASO2NOXNH3  30% reduction in all SO 2 /NO X /NH 3 domain-wide AMVOC  30% reduction in all Anthropogenic VOCs domain-wide ABVOC  30% reduction in all Biogenic VOCs domain- wide APRIC  30% reduction in all Primary Carbon domain-wide

35 Mammoth Cave (KY) Red line indicates the additional reductions in light extinction beyond 2018 OTB required to reach the Reasonable Progress Goals (Goal – OTB) Required reductions from 2002  79.7 Mm -1

36 Level 2&3 Sensitivity Acronyms GSO2ALL  30% reduction in all ground SO 2 domain-wide ESO2ALL  30% reduction in all point SO 2 domain-wide ESO2VCPP (CPP)  30% reduction in all VISTAS point coal-fired power plant SO 2 ESO2VNPP (NPP)  30% reduction in all VISTAS point non power plant SO 2 ESO2VOPP (OPP)  30% reduction in all VISTAS point non coal-fired power plant SO 2 ESO2nonV  30% reduction in all non VISTAS point SO 2 BCSO2  30% reduction in all SO 2 boundary conditions BCSO4  30% reduction in all SO 4 boundary conditions

37 Mammoth Cave (KY) Red line indicates the additional reductions in light extinction beyond 2018 OTB required to reach the Reasonable Progress Goals (Goal – OTB) Required reductions from 2002  79.7 Mm -1

38 Reasonable Progress for 2018 OTW? Yes No Maybe Undetermined Preliminary Results

39 Phase II Annual CMAQ Modeling

40 Phase II Modeling Approach Annual (12 month) CMAQ simulations to support regional haze SIP development –Will be modeling entire year of 2002 Emissions and Air Quality Modeling –Initial (completed) and Final (May 2005) AQ Modeling with “Actual” Baseyear Emissions Model Performance Evaluation –AQ Modeling with “Typical” Baseyear Emissions (April 2005) Same assumptions for Seasonal Distributions as Projected Future Year Emissions (Point Sources, Fires, etc.)  RRFs for SIP –AQ Modeling with Future Year (2018) Emissions (April 2005) On-the-Books (OTB) and On-the-Way (OTW) –AQ Modeling with Future Year (2018) Control Strategies (July 2005) –AQ Modeling with Future Year (2009) Emissions (May 2005) –AQ Modeling with Future Year (2009) Control Strategies (Aug. 2005) Final Report (delivery date December 2005)

41 Annual CMAQ Simulations Need to solve the Atmospheric Diffusion Equation for each species in each grid cell for each time step –(200 * 100 horizontal grid cells) x (19 vertical layers) x (100 species) x (4 time step/hour) x (24 hours/day) x (365 days/year) IN AN ANNUAL SIMULATION, NEED TO SOLVE OVER 1,330,000,000,000 PARTIAL DIFFERENTIAL EQUATIONS!!!! –Unix or Linux workstations (3.2 GHz ) –CPU time  ~ 4 months/simulation –CMAQ Inputs  3.0 TB –CMAQ Outputs  1.2 TB/simulation

42 Modeled Sulfate Bias

43 Modeled Nitrate Bias

44 Modeled Organics Bias

45 Modeled Elemental Carbon Bias

46 Modeled Soils Bias

47 Modeled Coarse Mass Bias

48 Performance Summary Sulfate and Elemental Carbon –Very good performance for all months Nitrate –Large over-predictions in winter –Updated ammonia monthly emission profiles (CMU model) Resulted in lower NH 3 emissions in winter (~60%) and much better nitrate performance Organics –Large under-predictions in summer –Updated CMAQ to include secondary organic aerosol (SOA) formation due to sesquiterpenes and polymerization (neither process currently accounted for in model) Resulted in much better organics performance Soils and Coarse Mass –Small contribution to light extinction due to small extinction coefficients

49 PM 2.5 and Ozone

50 PM 2.5 Modeling PM 2.5 NAAQS –15  g/m 3 (annual average over 3 years) –65  g/m 3 (24-hour average) Most states will use VISTAS modeling as starting point for PM 2.5 modeling –Annual modeling for 2002, 2009, and 2014 (?) PM 2.5 modeling collaboration between VISTAS states (e.g., AL and GA)

51 VISTAS 12 km ALGA 12 km

52 8-Hour Ozone Modeling 8-Hour Ozone NAAQS –Each monitor in an area must show the three year average of the fourth highest daily 8-hour ozone concentration to be 0.08 ppm or below –Average of three Design Values (2001, 2002, 2003) Some states will use VISTAS modeling as starting point for 8-hr ozone modeling Atlanta 8-hour ozone modeling –Will use ALGA 12 km modeling as starting point –Created new 4 km modeling domain –Modeling entire ozone season (05/20/02 – 09/20/02) –Modeling 2009 for attainment demonstration

53 VISTAS 12 km ALGA 12 km GA 4 km

54 GA Regional Sensitivities Sensitivity of ozone (ppb/tpd) and PM 2.5 (  g/m 3 /tpd) 1 Winter and 1 Summer Episode (  1 week) –Prefer 4 seasonal episodes with high summer ozone 10% Emission Reductions –NOx, VOCs, SO 2, NH 3, and primary PM 2.5 ALGA 12-km domain –4-km for summertime VOCs (?) Emission Regions –Atlanta, Macon, Columbus, Chattanooga, Floyd County 5 species * 4 episodes * 7 days/episode * 5 regions  700 modeled days

55 GA Point Source Sensitivities Sensitivity of ozone (ppb/tpd) and PM 2.5 (  g/m 3 /tpd) 1 Winter and 1 Summer Episode (  1 week) –Prefer 4 seasonal episodes with high ozone summer SCR (NO x ) and Scrubber (SO 2 ) Reductions –Discrete amounts (> 80%) ALGA 12-km domain –4-km preferred to capture plume structure Emission Locations –Bowen (SO 2 ), Scherer (NO x,SO 2 ), Branch (NO x, SO 2 ), Yates (NO x, SO 2 ), Wansley (SO 2 ), McDonough (NO x,SO 2 ), others (?) 10 scenarios * 4 episodes * 7 days/episode  280 modeled days

56 VISTAS Web Links UCR's website containing presentations, documents, and protocol associated with VISTAS Phase I emissions and air quality modeling (episodic simulations): –http://pah.cert.ucr.edu/vistas/docs.shtml UCR's website containing modeling results associated with VISTAS Phase I emissions and air quality modeling (episodic simulations): –http://pah.cert.ucr.edu/vistas/results.shtml UCR's website containing presentations, documents, and workplan associated with VISTAS Phase II emissions and air quality modeling (annual simulations): –http://pah.cert.ucr.edu/vistas/vistas2/docs.shtml UCR's website containing modeling results associated with VISTAS Phase II emissions and air quality modeling (annual simulations): –http://pah.cert.ucr.edu/vistas/vistas2/results.shtml

57 VISTAS Web Links (cont.) VISTAS website containing presentations and documents associated with VISTAS Workgroups (Data, Planning, and Technical Analysis): –http://www.vistas-sesarm.org/documents/index.asp BAMS website containing presentations and documents associated with episodic and annual MM5 meteorological modeling: –http://www.baronams.com/projects/VISTAS/ Georgia Tech's website containing presentations and documents associated with VISTAS emission sensitivities: –http://www.ce.gatech.edu/research/vistas/documents.htm Georgia Tech's website containing model results associated with VISTAS emission sensitivities: –http://www.ce.gatech.edu/research/vistas/products.htm

58 Jim Boylan Georgia Dept. of Natural Resources 4244 International Parkway, Suite 120 Atlanta, GA 30354 james_boylan@dnr.state.ga.us 404-362-4851 Contact Information

59 Questions?


Download ppt "Air Quality Modeling for VISTAS James W. Boylan, Ph.D. Georgia Department of Natural Resources (VISTAS Technical Lead for Air Quality Modeling) 2005 EPA."

Similar presentations


Ads by Google