Western Regional Air Partnership WRAP Regional Modeling Center

Western Regional Air Partnership (WRAP) Regional Modeling Center (RMC) Preliminary Fire Modeling Results Presented by: Ralph Morris WRAP Regional Modeling Center (RMC) rmorris@environcorp. com Presented at: Fire Emissions Joint Forum Meeting San Francisco, California June 3, 2003 Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 1

WRAP Regional Modeling Center (RMC) • University of California at Riverside (UCR) – Gail Tonnesen, Zion Wang, Jung Chien, etc. – Host RMC, CMAQ Modeling, Analysis • ENVIRON International Corporation – Ralph Morris, Gerry Mansell, Steve Lau, etc. – Interpretation of Results, MM 5 & REMSAD Modeling • UNC Carolina Environmental Program (MCNC) – SMOKE Emissions Modeling • WRAP Modeling Forum Co-Chairs – John Vimont (NPS), Mary Uhl (NM), Kevin Briggs (CO) • WRAP Technical Coordinators – Tom Moore and Lee Alter Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 2

Content of Today’s Talk • • Overview of WRAP Objectives Overview of Visibility Calculations WRAP § 309 SIP/TIP Modeling Approach CMAQ Model Performance Evaluation Use of Modeling Results to Project Future-Year Visibility Fire Management Practice Modeling Glide Path Slopes toward Natural Visibility Conditions Estimated 2018 Visibility Progress for § 309 Scenarios – Scenario #1: P 2 + Annex + BSM – Scenario #2: P 2 + Annex + OSM Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 3

WRAP Visibility Objectives • § 309 SIP/TIP due 2003 – 9 “Grand Canyon” states may opt-in (AZ, CA, CO, ID, NV, NM. OR, UT, and WY). – Focus on 16 Class I Areas on the Colorado Plateau • § 308 SIP/TIP due 2008 – 2000 -2004 visibility baseline – 2018 end of first planning period – Show progress toward natural visibility conditions by 2064 Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 4

Section 309 SIP/TIP Modeling Requirements • Demonstrate that SO 2 Annex Milestone control strategy is better than BART with Uncertainty • Analyze “significance” of Mobile Source and Road Dust at 16 Class I Areas • Estimate visibility improvements in 2018 due to § 309 All Control Strategy • Evaluate PM/NOx point source controls • Evaluate alternative fire management practices Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 5

WRAP § 309 Modeling Approach • 1996 Baseline Modeling Period • 36 -km Grid Covering Western US • SMOKE emissions modeling system using emissions provided by WRAP and EPA • Models-3 Community Multiscale Air Quality (CMAQ) modeling system • REMSAD model dropped from § 309 modeling due to time/resource constraints Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 6

WRAP CMAQ and REMSAD Modeling Domains Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 7

Components of Light Extinction • Light scattering and absorption – SO 4 sulfate, ammonium sulfate – NO 3 nitrate, ammonium nitrate – OC organic compound/organic matter – EC elemental carbon – PMF other fine particulates (<2. 5 ) – PMC coarse PM (2. 5 - 10 ) SO 4(NH 4)2 NO 3 NH 4 OC, OM, SOA Soot Soil PM 2. 5 -10 • NO 2 absorption considered a plume blight issue and not typically accounted for in regional haze assessments Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 8

Components of Light Extinction (continued) • Associated with each species is an “extinction coefficient” that converts concentration ( g/m 3) to light extinction (Mm-1) • Total visibility impairment is obtained as the sum of extinction due to each species: Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 9

WRAP Visibility Modeling (continued) • CMAQ 1996 Annual Runs – ~ 110 Gb of emission inputs – ~ 130 GB of other inputs – ~ 365 Gb of output • Initially annual simulations required 2 weeks – Multiprocessing allows runs to be completed in as little as 3 days • Challenge is processing 365 Gb of output into regulatory relevant results Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 10

WRAP Visibility Modeling (continued) • SMOKE emissions modeling becomes bottleneck – SMOKE QA/QC did not catch all errors in processing • Errors in treating holidays as weekdays • Many 2018 scenarios errors in allocating elevated sources dropped emissions • OSM vs BSM errors not caught – Interpretation of results requires matching runs in a consistent fashion (i. e. , with common errors) Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 11

WRAP CMAQ Model Performance Evaluation • ~30 IMPROVE sites in western US • Issues in matching monitored species with modeled species – Reconstructed Mass Equations – Actual Species • How to display results to convey performance • WRAP RMC website has 100 s of scatterplots and time series plots by site, by day, by month: http: //pah. cert. ucr. edu/rmc/models/index. shtml Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 12

1996 CMAQ Model Performance Issues • Nitrate overprediction bias especially in Winter and Spring/Fall – Ammonia emissions overstated under cold conditions • 2003 project to improve ammonia emissions – Deposition of ammonia and nitrate underestimated – June 2002 CMAQ release new heterogeneous nitrate formation • Exacerbated nitrate overprediction bias Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 13

1996 CMAQ Model Performance Issues • Some skill in sulfate estimates • EC, OC, and especially Soil highly scattered • Coarse Matter (CM) greatly underestimated – Missing local (subgrid-scale) impacts – Missing wind blown fugitive dust – 2003 project to develop wind blown dust inventory • Relatively better model performance is exhibited at sites on the Colorado Plateau and in the summer months when the Worst 20% days occur Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 14

Projecting Future-Year Visibility • Follow EPA draft guidance for projecting futureyear visibility (EPA, 2001 a, b, c) • Use model in a relative fashion to scale the current (1996) observed visibility for the Best 20% and Worst 20% days based on the ratio of the 2018 to 1996 modeling results – Relative Reductions Factors (RRFs) – Class I Area specific – Specific for each component of light extinction (SO 4, NO 3, EC, OC, Soil, and CM) Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 15

Projecting Future-Year Visibility • Accounting for missing fugitive dust emissions – No wind blown fugitive dust in inventory – Major component of observed Soil and CM – Model estimated RRFs for Soil and CM are in error • Set RRFs for Soil and CM to unity • RRF(Soil) = RRF(CM) = 1. 0 • Assumes 2018 Soil and CM concentrations are the same as 1996 Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 16

Glide Path Slope Values to Natural Visibility Conditions (NVC) • 2000 -2004 Observed Baseline Visibility Conditions (Anchors Glide Path Slope) – Worst 20% Days: Progress toward Natural Visibility Conditions in 2064 with Planning Periods ending at 2018, 2028, 2038, 2048, 2058, and 2064 – Best 20% Days: No Degradation in Visibility • Glide Path Slope Values assumes linear progress to Natural Visibility Conditions in 2064 Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 17

Preliminary Glide Path Slope Values to NVC • Use most current five-years of observed visibility to anchor Glide Path 2004 starting point for Worst 20% average visibility – 1995 -1999 used in preliminary analysis – Soon to be updated with 1997 -2001 data • Map Observed Visibility Conditions from Class I Areas with IMPROVE Monitoring to Nearby Similar Unmonitored Class I Areas • Use current EPA draft guidance for natural visibility conditions (NVC) for worst days (EPA, 2001) Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 18

Mapping of IMPROVE Data to Class I Areas Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 19

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 20

Preliminary Glide Path Estimates • Using Preliminary 1995 -1999 Observed Data – Will soon update to 1997 -2001 observations • Based on Current EPA Draft Guidance for Natural Visibility Conditions and f(RH) Values (EPA, 2001) – Revised Draft EPA Guidance expected soon • New f(RH) values are generally slightly lower • Have updated Glide Path Slope Value plots with new (2001) information Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 21

Projecting 2018 Visibility Improvements • Use relative changes in modeling results between 1996 and 2018 for average of Worst 20% (Best 20%) days to scale visibility baseline (1995 -1999 observed visibility) – Effects of changes in Soil and CM not accounted for [RRF(Soil) = RRF(CM) = 1. 0] • 2018 Projections for 2018 § 309 All Control Strategies Scenario Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 22

2018 § 309 All Control Strategy Scenarios#1&#2 • Area, Road Dust, Off-Road, On-Road Emissions – 2018 Base Conditions • Biogenic Emissions – 1996 Base Conditions • “Typical year” Wildfires Base Case • Point Sources – SO 2 Annex Milestones + Pollution Prevention) • Agricultural and Forest/Range Prescribed Fires – Scenario#1: Base Smoke Management (BSM) – Scenario#2: Optimal Smoke Management (OSM) • Example Emission Difference Plots for EC – Scenario#1 – Scenario#2 (BSM-OSM) Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 23

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 24

(BSM-OSM) Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 25

2018 Reasonable Progress Plots • 2018 Reasonable Progress Target Based on Preliminary Information – 1994 -1999 Observed Visibility – Preliminary f(RH) and Natural Conditions – Straight Line Projection from 2004 to 2064 • BSM Versus OSM Scenarios – Potential error in OSM scenario with daily emissions sometimes higher than BSM Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 26

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 27

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 28

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 29

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 30

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 31

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 32

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 33

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 34

Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 35

BSM Versus OSM Results • OSM Emissions Sometimes Higher Than BSM – Results in worsening in visibility if occurs during a day from the Worst 20% days • Need to Develop New OSM Emissions Inventory? – UNC/CEP emissions development delayed by lack of 2003 contract • Additional Fire Management Scenarios to be Modeled? Projects: /WRAP_RMC/Presents/ADEQ_Feb 062003. ppt 36