Evaluation and Intercomparison of N2O5 Chemistry in Two Versions of CMAQ

1
Evaluation and Intercomparison of N2O5 Chemistry
in Two Versions of CMAQ
Chao-Jung Chien, Gail S. Tonnesen, Bo Wang, Zion
S. Wang and Mohammad Omary
OVERVIEW The impact of tropospheric dinitrogen
pentoxide (N2O5) on the global distribution of
nitrogen oxides (NOx), ozone (O3) and hydroxyl
radical (OH) has been well recognized (Dentener
et al. JGR, v101, p.22869, 1996). Gaseous N2O5
reacts with water to form nitric acid (HNO3) both
in the gas phase and on surfaces (reactions ltR1gt
and ltR2gt). While the reaction of N2O5 in the gas
phase is relatively slow, with rate constant of
10-21 cm3 molecule-1 s-1, it has been shown the
heterogeneously hydrolyzed N2O5 could release up
to 70 of HNO3 into the gas phase (Wahner et al.,
JGR., v.103, p.31103, 1998). ltR1gt N2O5 H2O ?
2 HNO3 ltR2gt N2O5 H2O (aerosol) ? 2 HNO3 US
EPA has recently released its latest version of
the Community Multiscale Air Quality (CMAQ)
modeling system in June, 2002. Unlike the
previous version (released in March, 2001,
v.0301) which treats hydrolysis of N2O5
exclusively in the gas phase, the new release
(v.0602) includes heterogeneous conversion of
N2O5 to HNO3 on aerosol surfaces. In this
study, we evaluate CMAQ performance against
available ambient measurements, and compare the
contributing chemical processes using process
analysis for both versions of CMAQ. Other major
reactions (ltR3gt and ltR4gt) that convert NOx to
HNO3, the major sink for reactive NOx, are also
examined. ltR3gt OH NO2 ? HNO3 ltR4gt VOC
NO3 ? RO2 HNO3
RESULTS

• Difference in HNO3 production between conversion
  of N2O5 from New CMAQ (gas aerosol phases) and
  Old CMAQ (gas phase only). Plots are cumulative
  for one day in January and July.

• Difference plots in mixing ratio (ppbv) between
  new and old CMAQs for model species O3, HNO3,
  and aerosol nitrate (ANO3) in both January and
  July.

Jan.
July

• Time series plots for major chemical pathways
  contributing to HNO3 production.

Old
Old

• Model Evaluation with Ambient Database
• Database
• IMPROVE 49 stations in January and July, 1996
• CASTNET 25 stations in July, 1996 (limited data
  in January)
• CMAQ species
• ANO3 (IMPROVE, CASTNET), HNO3 (CASTNET), Total
  nitrates (CASTNET)
• Scatter plots
• AllSites_AllDays (Figures)
• Statistical analysis Regression (r-squared),
  Mean normalized bias (MNB) and error (MNE)

• Method
• Model description and comparisons
• Modified rate constant for reaction ltR1gt in new
  CMAQ.
• Implementation of heterogeneous conversion of
  N2O5 to nitric acid in new CMAQ (reaction ltR2gt)
• Based on studies by Dentener and Crutzen (JGR,
  v.98, p.7149, 1993)
• N2O5 H2O ? 2HNO3
• where k is the pseudo-first order reaction
  coefficient (s-1) r is the aerosol radius (cm)
  Dg (cm2 s-1) is the gas phase diffusion
  coefficient, calculated according to Fuller et
  al. (J. Phys. Chem., v.75, p.3679, 1969) ? is
  the reaction probability on surface particles (r
  0.1 for N2O5) v is the mean molecular speed (cm
  s-1) and A is the aerosol surface, (cm2
  aerosol)/(cm3 air).

New
New

• Summary of total HNO3 production for full domain
  on each day and for both versions of CMAQ

Model predictions vs. IMPROVE for Aerosol NO3 in
January
Model predictions vs. IMPROVE for Aerosol NO3 in
July

• IRR total HNO3 production summed for full domain
  and all days for both versions of CMAQ.

Model predictions vs. CASTNET for Gaseous HNO3 in
July
Model predictions vs. CASTNET for total nitrate
(ANO3HNO3) in July

• CMAQ Simulations
• WRAP domain
• 85 columns, 95 rows, 18 layers, 36km grid cells
  horizontally 68 variables
• Meteorology
• MM5 1996 simulation, processed with MCIP v.2
• Modeling period
• January 14-31
• July 18-31 (Julian dates 200-213)
• Spin-up periods Jan 1-13 July 1-17

• CONCLUSIONS
• New CMAQ shows more overproduction of nitrate
  than previous version in both January and July.
• Time-series plots for each reaction indicate
  diurnal changes for each processes dominant
  nighttime production of HNO3 from heterogeneous
  conversion of N2O5.
• Nocturnal heterogeneous conversion of N2O5 to
  HNO3 on aerosols accounts for up to 70 and 30
  of the diurnal HNO3 formation for New CMAQ in
  January and July, respectively (generally
  consistent with Dentener et al.).
• New CMAQ model performance is worse. It is
  uncertain whether the new N2O5 chemistry is wrong
  or if errors in other model inputs or ambient
  data are responsible for the poor model
  performance.

• Integrated reaction rate (IRR) analysis of two
  CMAQs
• Comparing process analysis outputs from CMAQs in
  mass based units (moles) on cumulative results
  (each day or 14 days total)
• Codes modified in new CMAQ (aero_subs.f) to
  extract information of HNO3 production from
  heterogeneous conversion of N2O5.

• Emissions processing
• SMOKE Mobil version D Area, Point, Biogenic
  version E
• CCTM modules
• QSSA with Process Analysis (Procan)
• Chem CB-IV with aqueous and aerosol
  extensionsCB4_AE2_AQ (old CMAQ) vs. CB4_AE3_AQ
  (new CMAQ)