CRGAQS: Revised CAMx Results

1
CRGAQSRevised CAMx Results

• Presentation to the
• Gorge Study Technical Team
• By
• ENVIRON International Corporation
• December 6, 2006

2
Todays Presentation

• Recap modeling performance issues
• Describe latest CAMx simulations
• Model and emission changes
• Performance for PM and light scattering
• Next Steps

3
Modeling Issues

• Episodes
• August 10-22, 2004
• November 4-18, 2004
• Identified issues from sensitivity runs
• Primary fine/coarse PM over predicted in both
  episodes
• Dominating modeled light scattering
• Windblown dust vs. fires?
• Nope construction agricultural fugitive dust

4
Modeling Issues

• OC over predicted in both episodes
• SOA is dominant in Aug (mainly biogenic)
• POA (EC) is dominant in Nov near Portland (wood
  smoke)
• Why is modeled scattering nearly zero in
  mid-November in the Mt Zion area?
• Easterly winds accelerate through Gorge and send
  Portland emissions offshore

5
Modeling Issues

• Why is modeled scattering so low at the eastern
  sites in November?
• No speciated data at Gorge study sites
• IMPROVE data on Nov 11 indicates dominance of NO3
  and OC
• OC, EC, and SO4 performance is good in east
• NO3 is under predicted in east
• We rely on modeled RH to generate the nitrate
  (complex process)
• We use observed RH to translate nitrate mass to
  nitrate scattering (simple process)

6
Revised Model Configuration

• SOA
• Historically under predicted by CMAQ and CAMx in
  the western U.S. (e.g., WRAP)
• Attributed to the biogenic component
• We employed a chemical improvement in CAMx for
  biogenic SOA
• Same as put into CMAQ for the RPOs
• Terpene 2-product mechanism higher yields,
  higher volatility
• Should reduce biogenic SOA

7
Revised Model Configuration

• Fugitive dust
• 2 SMOKE problems
• WRAP speciation profiles caused a double-counting
  of OR/WA dust estimates
• This also impacted OR/WA woodsmoke emissions
• No county-level canopy escape factor was
  applied (as developed and applied in WRAP)
• Both have been fixed and SMOKE re-run
• Significant coarse PM reductions
• Note WRAP has chosen to completely disregard
  primary coarse PM predictions

8
Performance Evaluation
9
August Performance Evaluation
Gorge Site Bscat
East Portland
10
August Performance Evaluation
Wishram
Bonneville
Mt. Zion
11
August Performance Statistics
12
August Performance Evaluation
13
August Performance Evaluation
14
Gorge Study vs. IMPROVEAt Mt. Zion
Organic and Elemental Carbon
15
Gorge Study vs. IMPROVEAt Mt. Zion
Sulfate and Nitrate
16
August Summary

• Overall good performance for
• Total PM2.5
• Light scattering
• Low SO4/NH4 and primary fine
• Questionable Gorge measurements
• NO3 insufficient to worry about
• High carbon
• SOA modification ineffective
• What is underlying cause?
• Model over predicts diurnal variation

17
November Performance Evaluation
Gorge Site Bscat
Portland
Eastern Gorge
18
November Performance Evaluation
Mt. Zion
Bonneville
Mt. Zion
19
November Performance Statistics
20
November Performance Evaluation
21
November Performance Evaluation
22
November Summary

• Generally high total PM2.5
• Dominated by carbon and primary fine
• Bifurcated performance for light scattering
• Over predicted in Portland area high carbon and
  primary fine
• Under predicted in eastern Gorge low SO4/NO3/NH4
• Need more humidity/clouds
• Generate more SO4
• Condense more NO3

23
Next Steps

• Recommendations
• Focus on August episode
• 2018 Case
• PSAT
• What-if scenarios
• Use model trends in relative sense to scale
  IMPROVE observations
• Relegate November episode
• Revisit with possible follow-on funding?