Data from a recent field campaign in Mexico City are used to evaluate the performance of the EPA Federal Reference Method for monitoring the ambient concentrations of NO2. Measurements of NO2 from standard chemiluminescence monitors equipped with molybdenum oxide converters are compared with those from Tunable Infrared Laser Differential Absorption Spectroscopy (TILDAS) and Differential Optical Absorption Spectroscopy (DOAS) instruments. A significant interference in the chemiluminescence measurement is shown to account for up to 50% of ambient NO2 concentration during afternoon hours. As expected, this interference correlates well with non-NOx reactive nitrogen species (NOz) as well as with ambient O3 concentrations, indicating a photochemical source for the interfering species. A combination of ambient gas phase nitric acid and alkyl and multifunctional alkyl nitrates is deduced to be the primary cause of the interference. Observations at four locations at varying proximities to emission sources indicate that the percentage contribution of HNO3 to the interference decreases with time as the air parcel ages. Alkyl and multifunctional alkyl nitrate concentrations are calculated to reach concentrations as high as several ppb inside the city, on par with the highest values previously observed in other urban locations. Averaged over the MCMA-2003 field campaign, the chemiluminescence monitor interference resulted in an average measured NO2 concentration up to 22% greater than that from co-located spectroscopic measurements. Thus, this interference has the potential to initiate regulatory action in areas that are close to non-attainment and may mislead atmospheric photochemical models used to assess control strategies for photochemical oxidants.
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 particle and ammonia data and to predict gas phase concentrations of nitric and hydrochloric acid. In general, the model is able to accurately predict the observed inorganic particle concentrations at all three sites. The agreement between the predicted and observed gas phase ammonia concentration is excellent. The NOz concentration calculated from the NOy, 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 particle phase observations during this period with the exception of the nitrate predictions after 10:00 a.m. (Central Daylight Time, CDT) on 9 April, where the model underpredicts the observations by, on average, 20%. This period had a low planetary boundary layer, very high particle concentrations, and higher than expected nitrogen dioxide concentrations. For periods when the particle chloride observations are consistently above the detection limit, the model is able to both accurately predict the particle chloride mass concentrations and provide well-constrained HCl (g) concentrations. The availability of gas-phase ammonia observations helps constrain the predicted HCl (g) concentrations. When the particles 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 particles 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.