Abstract. Partitioning of semi-volatile nitrate and ammonium between the gas and particulate phases is studied combining two thermodynamic models that explicitly include crustal elements and simulate both branches (deliquescence, efflorescence) of aerosol behavior and measurements taken near downtown Mexico City during a field campaign conducted in February–March, 2005. Overall, no significant differences between model predictions (within 30% of error) are observed for particulate ammonium (PM_2.5, PM₁). In cases of moderate to high RH (40–70%), mostly occurring during the 1st and 2nd daily sampling periods (06:00–10:00 h, 10:00–14:00 h, LST), 4 h PM_2.5 nitrate measurements are predicted within 30%. When RH drops below 30%, characteristic of the afternoon sampling periods (14:00-18:00 h), the efflorescence branch is most consistent with observed PM nitrate. Residual error analysis of these low RH cases suggest that aerosol nitrate loading or sulfate-to-nitrate molar ratio control phase behavior, hence the partitioning of semi-volatile PM_2.5 nitrate in gas and particulate phases. Finally, inclusion of crustal elements in the modeling framework reduces the error in predicted PM_2.5 ammonium by 25%. These findings, if generally applicable, can help improve air quality modeling in nitrate deficient environments.