Implementation of a Markov Chain Monte Carlo Method to inorganic aerosol modeling of observations from the MCMA-2003 Campaign. Part II: Model application to the CENICA, Pedregal and Santa Ana sites

F. M. San Martini (1) , E. J. Dunlea (2) , R. Volkamer (1) , T. B. Onasch (3) , J. T. Jayne (3) , M. R. Canagaratna (3) , D. R. Worsnop (3) , C. E. Kolb (3) , J. H. Shorter (3) , S. C. Herndon (3) , M. S. Zahniser (3) , D. Salcedo (4) , K. Dzepina (2, 5) , J. L. Jimenez (2, 5) , J. M. Ortega (1) , K. S. Johnson (1) , G. J. Mcrae (6) , L. T. Molina (1, 7) , M. J. Molina (1)

1 Division of Geological and Planetary Sciences [Pasadena]
2 CIRES - Cooperative Institute for Research in Environmental Sciences
3 Aerodyne Research Inc.
4 Centro de Investigaciones Quimicas
5 Department of Chemistry and Biochemistry [Boulder]
6 Department of Chemical Engineering
7 Molina Center on Energy and the Environment

A Markov Chain Monte Carlo model for integrating the observations of inorganic species with a thermodynamic equilibrium model was presented in Part I of this series. Using observations taken at three ground sites, i.e. a residential, industrial and rural site, during the MCMA-2003 campaign in Mexico City, the model is used to analyze the inorganic aerosol and ammonia data and predict gas phase concentrations of nitric and hydrochloric acid. In general the model is able to accurately predict the observed inorganic aerosol concentrations at all three sites. The agreement between the predicted and observed gas phase ammonia concentration is excellent. The NO_z concentration calculated from the NO_y, NO and NO₂ observations is of limited use in constraining the gas phase nitric acid concentration given the large uncertainties in this measure of nitric acid and additional reactive nitrogen species. Focusing on the acidic period of 9?11 April identified by Salcedo et al. (2006), the model accurately predicts the aerosol phase observations during this period with the exception of the nitrate predictions after 10:00 a.m. (CDT) on 9 April, where the model underpredicts the observations by, on average, 20%. For periods when the aerosol chloride observations are consistently above the detection limit, the model is able to both accurately predict the aerosol chloride predictions and provide well-constrained HCl (g) concentrations. When the aerosols are aqueous, the most likely concentrations of HCl (g) are in the sub-ppbv range. The most likely predicted concentration of HCl (g) was found to reach concentrations of order 10 ppbv if the aerosols are dry. Finally, the atmospheric relevance of HCl (g) is discussed in terms of its indicator properties for the possible influence of chlorine-mediated photochemistry in Mexico City.

Domaines: Océan, Atmosphère

F. M. San Martini, E. J. Dunlea, R. Volkamer, T. B. Onasch, J. T. Jayne, et al.. Implementation of a Markov Chain Monte Carlo Method to inorganic aerosol modeling of observations from the MCMA-2003 Campaign. Part II: Model application to the CENICA, Pedregal and Santa Ana sites. Atmospheric Chemistry and Physics Discussions, 2006, 6 (4), pp.5999-6040. ⟨hal-00303938⟩