WRAP 2004 Technical Work Elements Tom Moore March 24, 2004 Air Managers’ Committee §308/§309(g) Implementation Workgroup

2004 WRAP Technical Work Elements Topics Today Attribution of Haze Project TOC-sponsored Emissions Inventories EDMS collaborative effort – Emissions\Fire\Dust Forums Emissions Inventories’ Status – focus on fire EIs Regional Modeling Center Source Apportionment based on “interim” 2002 EIs AQ Modeling Forum-sponsored Causes of Haze Assessment Ambient Monitoring Forum-sponsored

2004 WRAP Technical Work Elements Start of technical and policy analysis path to Regional Haze Plans due by end of 2007 Technical analyses to understand: Baseline and 2064 natural conditions Reasonable Progress to national visibility goal Emissions reductions needed to achieve RP to 2018 Technical Activities - what is going on now?

2004 AoH Project Data Sources Source apportionment modeling simulations from the Regional Modeling Center Receptor-oriented source contribution analyses of aerosol and meteorological monitoring data from the Causes of Haze Assessment project Existing and refined emissions inventories from the Dust, Emissions, and Fire Forums Special-purpose source attribution studies such as BRAVO, et cetera EPA technical guidance documents and analyses Journal publications, and workshop/conference reports addressing emissions and visibility impairment

2004 AoH Project Deliverables Identify: Geographic source areas of emissions that contribute to impairment at each mandatory federal and tribal Class I area Mass and species distributions of emissions by source categories within each contributing geographic source area The amount of natural and manmade emissions affecting each Class I area

2004 AoH Project Deliverables Provide: Documentation of the assumptions, methods, and uncertainties used in the integrated analyses of modeling, monitoring, and emissions data. Succinct, clear summaries for policymakers, of the estimated areas and sources of impairment for each Class I area, including the associated uncertainty

AoH Project Schedule January – March Organizational meeting – March – workgroup page on WRAP website Develop scope of work for contractor support + hire April – June Review/discuss existing source attribution studies (BRAVO, et cetera) Contractor to identify data available for AoH project July – September Review/discuss work products from RMC, CoHA, and EIs Assign expert review topics

AoH Project Schedule October – December Continue review/discussion of work products from RMC, CoHA, and EIs Review/discuss draft reports from expert reviewers Review first draft of AoH report prepared by contractor January 2005 Publish final 2004 AoH report Make plan for subsequent workgroup activities

Emissions Data Management System Regional data center to identify existing, projected, and future control levels of emissions (more detail later this morning) Provide a “bucket” to store state and tribal EIs, as basis of Regional Haze Plans Complete EI for WRAP region Provide modeling input files Under construction now, on-line late

EDMS Project – Phase 1 Purpose Regional Haze Rule Implementation Performance Monitoring Region-Wide Emission Inventory Analysis of Collected Data Emissions Comparisons Emissions Trends Analysis

EDMS Project – Phase 1 Concepts GIS/Layers Need to add to map

2002 “Interim” Emissions Inventories Existing EI Data from §309 work Mobile (on-road and non-road) Road Dust (paved and unpaved) Improved EI Data Point and Area from 1999 NEI (CENRAP + WRAP) Ammonia Biogenics Windblown Dust EI Development Work Fire Mexico/Canada

Fire EI Tasks (task notes on following slides, resulting from discussions at 12/19/02 FEJF meeting) Task 1Task 2Task 3Task 4 Due Date4/048/041/055/05 Purpose 1) Model Evaluation 2) Test Apportionment WRAP Strategic Plan Phase I Apportionment Baseline Planning Apportionment 2018 Planning Apportionment Product 2002 Wx 2002 Rx 2018 Ag BSM Final 2002 Wx Final 2002 Rx Final 2002 Ag Representative EI for all fire types 2018 Representative EI(s) for all fire types

Task 1 Modeling Evaluation EI - Fire 2002 actual Wildfire EI – by April actual Rx fire EI (using NIFC and other state electronic data to supplement emissions estimates) – April actual Ag fire EI (may use 2018 BSM for now) Used for model performance evaluation – ambient monitoring data compared to model results Provides confidence in using model to characterize current and future years’ planning EIs Used to test geographic source apportionment – effects at each Class I area, tribal reservation, or other geographic area Provides modeling estimate of contributions from each upwind source jurisdiction

Task 2 Initial Modeling Apportionment EI - Fire Start with Phase I 2002 actual EI(s) 2002 actual Wildfire EI - 8/ actual Rx fire EI - 8/2004 (split natural versus anthropogenic) 2002 actual Ag fire EI - 8/2004 For all 3 types of fire: Used to complete Strategic Plan 2004 deliverable (geographic source apportionment) for TOC Attribution of Haze project Effects at each Class I area, tribal reservation, or other geographic area Modeling analysis of natural versus anthropogenic

Task 3 Planning Baseline Period EI - Fire Start/stay with Task EI format for consistency Use at least data, could be longer period, must be representative of regional haze baseline period Wildfire EI - 1/2005 Rx fire EI - 1/2005 Ag fire EI - 1/2005 To construct these EIs: Consult with states & tribes on smoke management programs Other considerations?

Task 4 Planning 2018 Projection Year EI - Fire Stay with Task 3 Baseline Planning 2002 EI format for consistency Base projections on predictable variables, may only provide ranges or scenarios of emissions Wildfire EI - 5/2005 Rx fire EI - 5/2005 Ag fire EI - 5/2005 To construct these EIs: Consult with states & tribes on smoke management programs Other considerations?

2004 Regional Modeling Center Workplan The Air Quality Modeling Forum has two major areas of activity planned for 2004: The Regional Modeling Center (RMC) will continue to operate and will implement many of the tools and improvements developed in The RMC will use the same team of contractors (UCR, ENVIRON, CEP) that have been used for the past 2 years The final work plan is being amended into the existing RMC contract. The second major area of activity is to initiate modeling for Class I areas in Alaska.

2004 Regional Modeling Center Workplan Major Elements 1. Project Administration year MM5 Modeling data – for now and later “Interim” Base Emissions Inventory processing 4. CMAQ Runs & Evaluation – based on modeling protocol 5. Source Apportionment (described later) 6. Natural vs. Anthropogenic Analysis 7. Windblown Fugitive Dust Model 8. Fire Sensitivity 9. Fire De Minimus 10. Alaska Modeling Lesser Elements 1. Emissions Speciation 2. Model Performance Software 3. Comparison of Alternate Models (AQ & meteorology) As needed: 1. Training

Gridded Dispersion Modeling - Source Apportionment Project Manager - Gail Tonnesen, University of California, Riverside Regional Modeling Center Team University of California, Riverside ENVIRON Corporation Carolina Environmental Programs at University of North Carolina

Motivation Need to understand which emissions sources contribute to haze and other pollutants. Europeans call these “Blame Matrices” Use this information to assist in developing control strategies.

Modeling Approaches Sensitivity Studies: Brute Force: Zero-in or Zero-out a single source. DDM Sensitivity – efficient but non-linear. Use tracers or “tagged species” to track mass from a source type: UCR and ENVIRON are implementing similar tracer algorithms in CMAQ and CAMx. Modeling back-trajectories. Chemical Mass Balance (CMB). Hybrid Approaches: OSAT Uses tracers to track O3 formation that was sensitive to VOC or NOx.

Tagged Species Approach Use “Tagged Species” tracers to track chemical transformations and the movement and chemical conversion of mass across domain. Add source type tracers for key species and for defined regions and source categories. Outputs 3-D fields showing transport of secondary species. Also outputs bar plots showing contributions at each receptor site.

Chemical Transformations Emissions are as NOx = NO + NO2 Use integrated reaction rates at each time step to update the tagged species: NOX  PAN NOX  Organic NO3 NOX  HNO3 HNO3  Aerosol NO3

Tagged Species for Nitrates NOX = reactive N family. = { NO, NO2, NO3, 2*N2O5, HONO, PNA} HNO3 PAN Organic NO3 Aerosol NO3

Traced Area: WRAP Modeling Domain Source Area Mapping File: Each state is distinguished by a unique number

Transport & Loss Terms Use CMAQ transport solvers for advection and dispersion of each tracer. Also update for mass export in cloud and aqueous chemistry algorithms. Update tagged species for emissions and deposition terms. Check for mass conservation at each step and adjust mass if needed. Halt if large errors.

Traced Source Tags TypesSource Category Notes ICON Initial Concentration BCON Boundary Concentration EmissionsMV_*Mobile sources from any state (on-road) BG_*Biogenic sources from any state RD_*Paved + unpaved road dust from any state NR_*Non-road sources from any state PN_*Point sources from any state AR_*Area sources from any state WF_*Wildland fire from any state AG_*Agricultural fire from any state RX_*Prescribed fire from any state ET_*Total emissions from any state *_WRAPAny type of source category emissions from WRAP domain OthersOTHERSAny sources other than all of the above

Contributions to Aerosol NO3 at Yosemite

Causes of Haze Assessment Project Project Manager – Mark Green, Desert Research Institute 4-year project – 2004 is year 2 CoHA Team – DRI + other analysts Userid = dri-coha Password = hazeyweb

CoHA Approach Determine causes of haze at WRAP and CENRAP Class I areas, tribal and selected CENRAP IMPROVE protocol sites 5

CoHA Study Data Began analysis of 1997 to 2002 IMPROVE and protocol database Primarily using IMPROVE and protocol sites with full speciation data in the study region (118 sites by December 2002) Using nationwide network of 158 sites (end of 2002) to establish continental and regional setting

CoHA not just computing statistics, but forming conclusions regarding the causes of the haze First complete a set of descriptive analyses, maps, and other graphics for aerosol composition, spatial and temporal variation, emissions, land use, topographic effects, transport patterns, local wind patterns etc Do episode analyses to determine likely causes of haze for various commonly and uncommonly occurring conditions Using above resources form conceptual models of causes of haze and assign quantitative number based on frequency of occurrence of conditions

Causes of Haze likely to be segregated by compound of interest, e.g. sulfate and by geographic area - by source type as possible Example: Sulfate causes 50% of aerosol haze at Area A - 60% of which is generated within the WRAP area, mainly in the states of B,C,and D, 20% is transported into the RPO from states to the east of WRAP and mainly in summer, and 20% from other countries (mostly Country F). Based upon emissions inventory, it is estimated that 80% of the sulfate haze is due to source type G. Nitrate is X % of the haze, 50% of which … Carbon, coarse mass - probably more difficult

The Causes of Haze web site is online now in a DRAFT, password protected form: Username: dri-coha Password: hazeyweb 8 Much of the web site is a shell ready to receive data and causes of haze information that we generate

Aerosol Descriptive Analysis Provides answers to the questions: For the years , how many measurements are available for the site in each month of each year, and what are the contributions of the major aerosol components to light extinction in each month of each year? What is the overall average light extinction at the site, and what are the contributions of the major aerosol components to the light extinction? What are the light extinction contributions by the major aerosol components for best, worst and average days and how do they compare? What percentage of the sampling days are the worst days in each month & how variable are the chemical components?

Sample Aerosol Description Page Overall average light extinction and contributions of major aerosol chemical components to light extinction Average contributions of major aerosol chemical components to light extinction in 20% best, middle 60% and 20% worst days Percentage of sampling days that are 20% worst days in each month Average contributions of major aerosol chemical components to light extinction during 20% worst days in each month

Meteorological & Emissions Descriptive Analysis Archived monitoring network locations, climate, emissions, wildfires, census, political, physical, and image databases Information from these databases are helping us build conceptual models and answer descriptive analysis questions by visualizing data (e.g. map emissions densities) Assist us in the general and detailed description of the meteorological setting of each site Creating maps of emissions surrounding each site at two scales: 2 km and 20 km- Regional emission maps to be added Include table of surrounding point sources ranked by distance and emission rate

20 Km terrain

2 Km terrain

Three years ( ), three heights (10, 500, 1500m), every three hours, 8 days back HYSPLIT v4.6 model calculations done for all sites Trajectory output processed and stored in database Trajectory tool developed to produce ASCII summary files and convert trajectories into shape files Generate summary maps Generate monthly and annual residence time maps, 20% best, 20% worst extinction, conditional probability Finalizing process to generate all maps, all sites in one batch- should be done late February Trajectory Analysis Status

Episode Analysis Use combination of back trajectory, synoptic, mesoscale meteorological analysis, aerosol and emissions data to conceptually understand single site and regional or sub-regional episodes of high aerosol component concentrations Systematic survey of episodes from the 1997 to 2002 IMPROVE database

“Hazagon” Analysis The hazagon provides a way to visualize speciated extinction for those sites in the 20% worst category 25

Future Phases Evaluation of EDAS wind field used for back trajectory analysis – when adequate, when misleading- possible use of MM5 or diagnostic wind fields for trajectory analysis for some sites Mesoscale meteorological analysis – trajectory analysis? Needed for sites in complex/coastal setting affected by mesoscale source areas Triangulation of back trajectories for worst case days to better identify source areas Regression analysis of back trajectories, aerosol data for quantitative attribution to regions- Trajectory Mass Balance Regression Refinement of conceptual models