Shaocai Yu*, Robin Dennis* , Brian Eder* , Shawn Roselle* ,

1
Can the thermodynamic model and 3-D air quality
model predict the aerosol NO3- reasonably?

•  4.      Conclusion
• ISORROPIA and AIM have been evaluated by
  comparing the modeled partitioning of TNO3 and
  TNH4 between gas and aerosol phases with
  observations. At the Atlanta site, both models
  could reproduce most of observed NH4 and
  HNO3within a factor of 2. However, both models
  cannot reproduce most of observed NO3- and NH3
  within factor of 2. At the Clinton site, both
  models performed a little better on NO3- than at
  the Atlanta site. There are different reasons for
  the model unable to reproduce NO3- reasonably.
  Sensitivity tests show that both models have
  similar responses in the predicted aerosol NO3-
  to the possible errors in SO42- and TNH4. This
  analysis indicates that errors in TNH4 are more
  critical than errors in SO42- to prediction of
  NO3-. Regardless, the 3-D model performance on
  SO42- and TNH4 needs to be quit good and better
  than current daily performance, before the 3-D
  air quality model can predict aerosol NO3-
  reasonably although it can predict TNO3
  reasonably.

•  1.   INTRODUCTION    
• Studying the behaviors of nitrate is one of most
  intriguing aspects of atmospheric aerosols
  because particulate nitrate concentrations depend
  not only on the amount of nitric acid, but also
  on availability of ammonia, sulfate
  concentrations, temperature and relative
  humidity. Particulate nitrate is produced
  partially or predominantly from the equilibrium
  reaction between two gas-phase species, HNO3 and
  NH3. Several inorganic thermodynamic models have
  been developed to partition the semi-volatile
  species between gas and aerosol phases during the
  past two decades. It has been postulated and
  confirmed by ambient measurements that the
  sulfate/nitrate/ammonium aerosol constituents
  should be in thermodynamic equilibrium with the
  local gas phase (Nenes et al., 1999 Ansari and
  Pandis, 2000 Moya et al., 2001). It is still one
  of the most challenging tasks to partition the
  semi-volatile inorganic aerosol components
  between the gas and aerosol phases correctly,
  especially when the thermodynamic models are
  incorporated in a 3-D air quality model.
• 2. MODELING AEROSOL NITRATE THERMODYNAMICS AND
  OBSERVATIONAL DATASETS
• 2.1. Thermodynamic Models
• SO42-, TNH4 (NH4 NH3) and TNO3 (NO3-HNO3 ) as
  input to
• ISORROPIA the optimal solution of the
  thermodynamic equations and precalculated tables,
  whenever possible, to speed up (Nenes et al.,
  1999).
• AIM2-Model II (Clegg et al., 1998) a
  theoretically complete and accurate phase
  equilibrium model not apply any simplifying
  assumptions.
• 2.2. Observational Dataset
• At the Atlanta site PM2.5 SO42-, NO3-, and NH4
  were measured with a 5-minute sampling (8/18 to
  9/1, 1999) (Weber et al., 2003). NH3 (g) and
  HNO3 (g) were measured with a time resolution of
  15 and 9 minutes, respectively. Temperature and
  RH were determined. Total 325 data points.
• At the Clinton Horticultural Crop Research
  Station, NC, PM2.5 NH4, NO3- and SO42-, and gas
  NH3 and HNO3 were measured with 12-hour
  resolution (1/20 to 11/2, 1999). Temperature and
  RH were provided by State Climate Office of NC.

Shaocai Yu, Robin Dennis, Brian Eder, Shawn
Roselle, Athanasios Nenes, John
Walker Atmospheric Sciences Modeling
Division, National Exposure Research
Laboratory, Air Pollution Prevention and
Control Division, National Risk Management
Research Laboratory, U.S. EPA, NC 27711
Schools of Earth and Atmospheric Sciences and
Chemical and Biomolecular Engineering Georgia
Institute of Technology, Atlanta Georgia 30332
On assignment from the National Oceanic and
Atmospheric Administration, U.S. Department of
Commerce
 

• 3. Results and discussions
• 3.1. Test of thermodynamic models with
  observational data
• Figure 1 and Table 1 for Atlanta site 94 and
  96 of the NH4 predictions are within a factor
  of 1.5 for ISORROPIA and AIM2, respectively.
  HNO3 86 (ISORROPIA) and 87 (AIM2) within a
  factor of 1.5. However, both models cannot
  reproduce most of observed aerosol NO3- and gas
  NH3 ( NO3- 32 (ISORROPIA) and 48 (AIM2)
  within a factor of 2, and NH3 25 (ISORROPIA)
  and 51 (AIM2) within a factor of 2.
• Figures 2 and 3 overprediction are associated
  with low temperature (T), high RH and
  sulfate-poor conditions (TNH4/SO42-gt2.0)
  underpredictions are associated with high T, low
  RH and sulfate-rich (TNH4/SO42-lt2.0).
• Figure 4 both models reproduced observed NH3
  concentrations very well (95 (ISORROPIA) and 97
  (AIM2) within a factor of 1.5) at Clinton site.
• Figure 5 Most of the cases at the Clinton site
  are representative of very sulfate-poor
  conditions.
• The possible reasons for poor performance
• (1) a dynamic instead of an equilibrium model
  may be more suitable for these cases. Moya et
  al., (2001) a dynamic instead of an equilibrium
  model was good for the cases with high T and low
  RH values.
• (2) Models are not able to accurately simulate
  such cases for the conditions encountered Ansari
  and Pandis (2000) metastable state assumption
  predicted 11 higher of NO3- than stable state
  assumption.
• (3) Other ions (such as Na, Cl-, Ca2 and
  Mg2) significant contributions to aerosol
  components and their effects not considered
  Coarse mode effects.
• (4) Other mechanisms such as absorption by
  carbonaceous aerosol instead of thermodynamic
  equilibrium produce aerosol NO3- Middlebrook et
  al. (2002) results at Atlanta site.
• (5) There are significant errors in observations
  of other important aerosol components (such as
  SO42-) and TNH4. See section 3.3

• 3.2. Situation about simulations of SO42-,TNH4,
  NO3- by 3-D air quality models
• Table 2 and Figures 6 and 7 for CMAQ
  reproduced 46-79 and 68-94 of SO42- within a
  factor of 1.5 and 2, respectively reproduced
  39-72 and 61-86 of NH4 within a factor of 1.5
  and 2, respectively reproduced 11-31 and 17-51
  of NO3- within a factor of 1.5 and 2,
  respectively much lower than TNO3 (34-59 and
  66-78 within a factor of 1.5 and 2,
  respectively).
• 3.3. Effects of errors in SO42-, TNH4, T and RH
  on predicting aerosol NO3-
• Test dataset of total 163 data points both
  ISRROPIA and AIM2 predict the existence of
  aerosol NO3-, based on observational data at the
  Atlanta site.
• Base-case results The prediction results of each
  thermodynamic model for the test dataset before
  introduced errors.
• Figure 8 and Table 3
• both ISORROPIA and AIM2 have similar responses in
  the predicted NO3- to the possible errors in
  SO42- and TNH4 NO3- predictions by ISORROPIA are
  modestly more sensitive to the errors in SO42-
  and TNH4 than those by AIM2.
• Conditions with -50 errors in TNH4 including
  cases 5, 6 and 7, both ISORROPIA and AIM
  underpredict almost all NO3- by more than a
  factor of 2.
• Conditions with 50 error in SO42- (case 1), or
  50 error in TNH4 (case 2), or 50 error in
  SO42- and 50 error in TNH4 (case 8), both model
  cannot reproduce most of NO3- within a factor of
  2 (percentagelt40).
• Conditions with the case 3 (50 errors in both
  SO42- and TNH4) and case 4 (-50 errors in
  SO42-) relative less effects on the prediction
  of NO3-. This is because of compensation error
  from both SO42- and TNH4.
• Figure 9 and Table 3 Responses of the NO3
  predictions are less sensitive to errors in T and
  somewhat less sensitive to errors in RH. ?20
  errors in T and RH result in both models not
  being able to reproduce most of NO3- within a
  factor of 1.5 (percentagelt42)

Acknowledgements The authors wish to thank other
members at ASMD of EPA for their contributions to
the 2003 release version of EPA Models-3/CMAQ
during the development and evaluation. This work
has been subjected to US Environmental Protection
Agency peer review and approved for publication.
Mention of trade names or commercial products
does not constitute endorsement or recommendation
for use. REFERENCES Clegg, S.L., Brimblecombe,
P., and A.S. Wexler, 1998 J. Phys. Chem. A,
102,2155-2171. Nenes, A., Pilinis, C., and S.N.
Pandis, 1999 Atmos. Environ., 33, 1553-1560.
Moya, M., Ansari, A.S., Pandis, S.N., 2001.
Atmospheric Environment, 35, 1791-1804,
2001. Weber, R.J., et al., 2003 J. Geophys.
Res., 108, 8421, doi10.1029/2001JD001220. Ansari,
A.S., and S.N. Pandis, 2000. Atmospheric
Environment, 34, 157-168. Meng, Z., and J.H.
Seinfeld, 1996. Atmospheric Environment 30,
2889-2900. Middlebrook, A.M., et al., 2002. A
comparison of particle mass spectrometers during
the 1999 Atlanta Supersite Project. J. Geophys.
Res.,