CMAQCAMx MPE for VISTAS 2002 Episode

1
Updates to CMAQ Secondary Organic Aerosol Module
(SOA) ENVIRON, Alpine Geophysics, and UC
Riverside
VISTAS TAWG Meeting Asheville, North
Carolina August 17-18, 2005
2
Content CMAQ SOA Module Updates

• 2004 Organic Carbon (OC) performance problems
  with standard CMAQ V4.3/V4.4
• Nov 04 - Jan 05 Identify missing processes in
  CMAQ SOA module and implement prototype SOAmods3
  update
• Feb 05 May 05 Peer review, further updates
  and testing.
• Jun 05 Aug 05 Implement comments and
  evaluate
• CMB and OC/BC estimates of SOA comparisons
• Update OMC 1.8 x OC factor (from 1.4)
• August 2005 Need decision on version of CMAQ
  for Base F modeling

3
SOA Module Updates Fall 2004

• Initial 2002 CMAQ V4.3 Simulation Exhibited Large
  (-80) Summer Organic Matter Carbon (OMC)
  Underestimation Tendency OMC 1.4 OC

OMC Monthly Fractional Bias -- IMPROVE VISTAS
4
Missing SOA Processes in CMAQ/CAMx

• Polymerization Part of SOA becomes nonvolatile
• Reactivity CGs are reactive VOCs that should
  participate in photochemistry already accounted
  for?
• Uncertain SOA Parameters CalTech chamber data
  used to fit CMAQ SOA parameters found to have
  artifacts anything better?
• Sesquiterpenes Not included, known important SOA
• Isoprene Not included, high in southeastern US
• Biogenic OVOC Also SOA precursor
• Acid Catalyzed reactions May be important for
  some SOA formation (e.g., Isoprene) mechanism
  needed?
• SOAmods3 vs. SOAmods4 Updates

5
SOA Module Updates January 2005

• Use latest CMAQ V4.4 as Starting Point
• CMAQ V4.4 exacerbated (-110) the CMAQ V4.3
  summer OMC underestimation bias (-80)
• Updated CMAQ SOA Module (SOAmods3)
• SOA from Sesquiterpenes
• SOA from Isoprene
• Polymerization of volatile SOA
• Keep same CMAQ V4.4 Inputs
• Performed CMAQ SOAmods3 2002 Annual Simulation
  (Base E)
• Improved OMC model performance

6
SOAmods3 Updates -- Red New Processes
G_ISP
Y1
O
O
ISOP
ISPD
ISPRXN
SGTOT_ISP
A_ISP
ASOC3
P1
PAR OLE ALD2
G_TRP1
SGTOT _TRP1
Biogenic emissions
Y3
A_TRP1
EF1
G_TRP2
AORGB
Y4
O
SGTOT _TRP2
TERP
TERPRXN
A_TRP2
EF2 ( EEG1 Y2 EEA1 )
ASOC2
OVOC
FPOLY
ASOC1
All SOAs except SOC1 SOC2
Polymerization
PAR OLE NR
O Oxidation reaction with OH, O3, and
NO3 ISPRXN the amount of reacted ISPD with
oxidants TERPRXN the amount of reacted TERP with
oxidants Dotted box not implemented in the
initial test
7
SOAmods3 Updates OMC Performance
CMAQ V4.4 versus CMAQ SOAmods3 OMC 1.4 x OC --
Monthly Model Performance
8
SOAmods Update Jan-Jun 2005

• SOAmods3 Update Journal Paper submitted February
  2005 to Atmospheric Environment for Publication
  in Special CMAS Edition
• Reviewer comments received June 11, 2005
• Manuscript updated and re-submitted August 2005
• SOAmods4 Updates
• Add OVOC emissions to CMAQ Inputs
• Add SOA from OVOC
• Ratio Sesquiterpene emissions off of TERPB and
  OVOC emissions

9
SOAmods4 SOAmods3 Updates New Updates
Y1
P1
O
O
ISOP
ISPD
ISPRXN
SGTOT_ISP
A_ISP
ASOC3
PAR OLE ALD2
SGTOT _TRP1
Y3
A_TRP1
Biogenic emissions
TERPRXN
AORGB
SGTOT _TRP2
Y4
A_TRP2
O
EF1
TERP
TERP
EF2 (EEG1 Y2 EEA1)
SGTOT _SQT
ASOC2
EF3 (EEG1 Y2 EEA1)
SGTOT _OVOC
EF4 (EEG2 EEA2)
P2
OVOC
ASOC4
A_OVOC
PAR OLE NR
FPOLY
ASOC1
All SOAs except SOC1 SOC2
Polymerization
10
Evaluation of SOAmods Updates

• CMAQ V4.4, SOAmods3 and SOAmods4
• SOAmods4 July 2002 Only
• OMC 1.4 x OC and OMC 1.8 x OC
• New OMC/OC 1.8 factor has been generally adopted
• SEARCH Network CMB C14 Analysis
• CMB estimates primary OMC (POM) contribution, SOA
  contribution inferred from differences
  (SOAOMC-POC)
• C14 separates new (biomass burning SOA from
  biogenics) and fossil (combustion) OMC
• Less useful as model combines all primary
  emissions
• MRPO OMC CMB analysis for Seney NWR, Michigan
• MRPO Urban OMC analysis in progress

11
Evaluation of SOAmods Updates

• SOAmods4 versus SOAmods3 Differences
• Sesquiterpene emissions ratio off of TERPB and
  OVOC instead of just TERPB
• Small increases in biogenic SOA
• SOA pathway directly from OVOC very small
• SOAmods4 Updates from EPA Review
• EPA identified need to include recalculation of
  number and surface area (modes) when adding
  species
• Has small effect on OMC and other PM estimates

12
July 2002 VISTAS/IMPROVE OMC1.4 x OC CMAQ
SOAmods3 Effects of including surface area/number
(blue) in SOAmods3 as recommended by EPA Very
small effect across all species
13
July 2002 VISTAS/IMPROVE OMC1.4 x OC CMAQ
SOAmods3 vs. CMAQ SOAmods4 Fractional bias
changes from 5 to 12
14
July 2002 VISTAS/IMPROVE OMC1.8 x OC CMAQ
SOAmods3 vs. CMAQ SOAmods4 Fractional bias
changes form -28 to -11
15
OMC 1.8 x OC July 2002 IMPROVE CMAQ V4.4,
SOAmods3 SOAmods4 Model Performance CMAQ
SOAmods4 best performing model for OMC across
eastern U.S. domains, worst in western U.S. (WRAP)
16
SEARCH SOA Estimates

• Chemical Mass Balance (CMB) receptor model used
  to estimate Primary Organic Matter (POM)
  concentrations at 8 SEARCH sites
• CMB estimates of SOA obtained by difference
  between measured OMC and CMB estimated POM
• Acknowledge M. Zheng, E. Edgerton, J. Jansen,
  etc.
• Use of OC/BC ratios to estimate SOA
• Site specific analysis with stratification
• Acknowledge R. Saylor, E. Edgerton, B. Hartsell,
  J. Jansen, etc.
• Model Model Comparisons

17
CMB vs. CMAQ V4.4 SOAmods3 POM/SOA January
(OMC 1.8 x OC)
Atlanta (JST) and Yorkville (YRK), Georgia and
Birmingham (BHM) and Centreville (CTR), Alabama
18
CMB vs. CMAQ V4.4 SOAmods3 POM/SOA January (OMC
1.8 x OC)
Oak Grove (OAK) and Gulfport, Mississippi (GFP),
and Pensacola (OLF and PNS) Florida
19
CMB vs. CMAQ V4.4, SOAmods3 SOAmods4 POM/SOA
July (OMC 1.8 x OC)
CMAQ V4.4 severely underestimates CMB estimated
SOA, whereas SOAmods has slight overestimation
20
CMB vs. CMAQ V4.4, SOAmods3 SOAmods4 POM/SOA
July (OMC 1.8 x OC)
CMAQ V4.4 large SOA underestimation relative CMB,
SOAmods much closer
21
July 2002 VISTAS/SEARCH Monthly Average SOA
Estimates CMAQ vs. CMB Fractional Bias/Error V4.4
-121/121 SOAmods3 22/34 SOAmods4
32/43
22
Seney, MI SOA/POM CMB CMAQ SOAmods3
Estimates Monthly Avg. SOA/POM Splits (OMC 1.8
x OC)
23
Seney NWR, Michigan, CMB vs. CMAQ V4.4, SOAmods3
SOAmods4 July 2002
SOA 3.2 µg/m3 0.8 µg/m3 2.9 µg/m3
4.0 µg/m3
24
SEACH OC/BC Ratio SOA EstimatesJuly 2002 OMC
1.8 x OC
CMAQ V4.4 SOA outside OC/BC ratio SOA uncertainty
estimates both SOAmods within uncertainty
25
Conclusions SOAmods

• CMAQ SOAmods much improved OMC performance in
  summer over standard CMAQ V4.4
• Winter performance mixed
• SOA estimates from CMB and OC/BC consistent with
  SOAmods updates but not with CMAQ V4.4
• SOAmods4 slightly higher OMC with slight
  degradation in OMC performance
• Need decision on 2002/2018 Base F modeling CMAQ
  V4.4 SOAmods3 or SOAmods4
• Also include SULF correction identified by AER/SCS