Comprehensive Particulate Matter Modeling: A One Atmosphere Approach

1
Emissions Sensitivity Modeling with DDM-AERO
James W. Boylan Georgia Department of Natural
Resources (VISTAS Technical Lead for Air Quality
Modeling) VISTAS STAD Meeting Charleston,
SC February 5, 2004
2
Outline

• Modeling Overview
• DDM-AERO Phase I Modeling
• Model Development and Testing
• CMAQ Speed-Up
• Schedule for Deliverables
• DDM-AERO Phase II and III
• Modeling Plans
• Schedule for Deliverables

3
DDM-AERO Modeling Overview

• Sole Source Justification with Georgia Tech
• GT has already installed DDM (gas-phase) into
  CMAQ and DDM-AERO in URM-1ATM
• Install DDM-AERO into latest version of CMAQ
• Allow evaluation of at least twice the number of
  sensitivities compared to brute force technique
• Apply CMAQ with DDM-AERO to VISTAS three Phase I
  episodes using 2015 emissions to help with
  control strategy development

4
DDM-AERO Work Schedule

• Phase I DDM-AERO Model Development and Initial
  Testing
• December 2003 February 2004
• Phase II Additional Testing and Detailed
  Evaluation
• March 2004 - April 2004
• Phase III Application of DDM-AERO to VISTAS
  Phase I episodes
• May 2004 - October 2004

5
DDM-AERO Phase I

• Code Development and Installation
• Inert PM Species (EC, Soils)
• Inorganic PM Species (SO4, NO3, NH4)
• Organic Species (OC)
• Initial Testing using In-House Modeling
• Georgias Fall-Line Air Quality Study (FAQS)
• August 10 - 18, 2000 episode
• Test Domain 20 x 20 grid cells (36 km)
• Expanded Full FAQS Domain - Eastern US (36 km)
• Qualitative Analysis

6
Sulfate Sensitivity to SO2 Emissions
Brute Force DDM
7
Organic Aerosol Sensitivity to Terpene Emissions
Brute Force DDM
8
DDM-AERO Phase I (cont.)

• Test Feasibility to Improve CMAQ Efficiency
• Modeling time step calculations are based on wind
  speeds
• CMAQ uses the maximum wind speed in the domain to
  determines the time step for every grid cell in
  the domain
• Evaluate variable time step algorithm
• Cells with lower wind speeds have larger time
  steps and use less CPU time
• Potentially a factor of 1.6 to 2.0 speed-up
• Allow more sensitivities to be tested within
  VISTAS schedule
• Can potentially apply this algorithm to reduce
  CPU time for annual simulations
• Recommendations to VISTAS in March 04

9
Surface Wind Fields on July 10 (12 km)
10
DDM-AERO Phase II

• Test and evaluate DDM-AERO code on VISTAS Phase I
  episodes (winter/summer)
• Development of DDM Performance Evaluation
  Software
• Quantitative Analysis (Bias, Error, r2, etc.)
• Determine limits of linearity for each
  pollutant
• Examine combined sensitivity runs
• SNOx SSO2 S(NOxSO2)

?
11
DDM-AERO Phase II (cont.)

• If needed, modify DDM code to run with
  alternative model configuration
• CB-IV, CB4-2002, SAPRC-99, CMAQ-AIM, etc.
• If recommended, install new variable time step
  algorithm into CMAQ
• Draft report due April 2004

12
DDM Phase III

• Apply DDM-AERO to VISTAS Phase I episodes with
  2015 OTB/OTW emissions
• Sensitivities of the individual components of PM
  and light extinction to emissions from
• Different emission source categories
• Point SO2, Area SO2, Point NOx, Mobile NOx, Area
  NH3, etc.
• Different sub-domains and boundary conditions
• FL, MS, AL, GA, SC, NC, TN, KY, WV, VA, MANE-VU,
  MRPO, CENRAP
• Report and Recommendations due October 2004

13
Emissions, Meteorological, Air Quality Modeling
Deliverables
Draft 1/16/04
Jan-Mar 2004 Define inv growth and control
assumptions
Jan-Jun 2004 Define BART sources
Jun 2004 Identify BART controls
June 2005 Economic Analyses
Apr 2004 Draft 2015 National Inv
Oct-Dec 2004 Control Strategy Inventories
Jan 2004 Revised 2002 VISTAS Em Inv
Sep 2004 Revised 2002 National Inv
Sep 2004 Typical 2002 Modeling Inv
Feb 2004 Em Modeling QA Fill Gaps
Oct 2004 Revised 2015 Em Inv
Feb 2004 MM5 Met runs 6 mo 2002
Jan 2004 Met modeling protocol
Sept 2004 MM5 Met Final Report
Jan-Jun 2005 2015 Control Strategy Runs
Dec 2004 Revised 2002 Base Run (model performance)
Dec 2004 Typical 2002 Run (compare to 2015
runs)
Dec 04 ? 2015 Base Run
Feb 2004 AQ modeling protocol
Jan 2004 AQ Phase I wrapup
Mar-Sep 2004 Annual 2002 CMAQ model performance
Jan 2005 Phase II 2015 Sensitivity Runs
July-Dec 2005 Observations Conclusions Recommenda
tions
EPA- approved Modeling Protocol
May-Oct 2004 2015 Emissions Sensitivity Runs
Apr 2004 DDM in CMAQ
Apr 2004 CARTselect episodes
Aug 2004 Natural Background and Reasonable
Progress Goals
State Regulatory Activities
Jan 2005 Interim Future Year Model Runs
Dec 2004 Interim Future Year Inventories