References

The NO-NO2 system was analyzed in different chemical regimes/air masses based on observations of reac- tive nitrogen species and peroxy radicals made during the in- tensive field campaign MIRAGE-Mex (4 to 29 March 2006). The air masses were categorized into 5 groups based on com- binations of macroscopic observations, geographical loca- tion, meteorological parameters, models, and observations of trace gases: boundary layer (labeled as "BL"), biomass burning ("BB"), free troposphere (continental, "FTCO" and marine, "FTMA"), and Tula industrial complex ("TIC"). In general, NO2/NO ratios in different air masses are near pho- tostationary state. Analysis of this ratio can be useful for testing current understanding of tropospheric chemistry. The ozone production efficiency (OPE) for the 5 air mass cate- gories ranged from 4.5 (TIC) to 8.5 (FTMA), consistent with photochemical aging of air masses exiting the Mexico City Metropolitan Area.


Introduction
During March of 2006, the Megacities Impact on Regional and Global Environment: Mexico (MIRAGE-Mex) field campaign took place in the region of the Mexico City Metropolitan Area (MCMA) including the Gulf of Mexico.This field campaign was designed to examine the chemical and physical transformations of gases and aerosols in the polluted outflow from MCMA and to assess the current and future impacts of these exported pollutants on regional and global air quality, ecosystems, and climate.MCMA is a megacity of about 20 million people residing in an area of 1500 km 2 , surrounded by mountains and at an elevation of 2.2 km above sea level (asl).The air quality in MCMA is affected by strong anthropogenic sources of NO x (NO+NO 2 ) and volatile organic compounds (VOCs) in conjunction with high solar irradiance facilitating photochemistry (Raga et al., 2001).In the 1990's, hourly averaged ozone concentrations in MCMA exceeded the Mexican national standard of 110 ppbv for much of the year (Raga and Raga, 2000).As essential ingredients for the formation of O 3 , the reactive nitrogen species (NO x ) are emitted to the atmosphere mainly in the form of NO from road traffic (Soltic and Weilenmann, 2003).NO and NO 2 are interconverted rapidly through following reactions: NO 2 +hν(λ<420 nm)(+O 2 )→NO+O 3 (R2) where RO 2 is any organic peroxy radical including CH 3 O 2 , and RO is the corresponding alkoxy radical.The NO 2 /NO ratio can be formulated as: where k 1 , k 3 , and k 4 are reaction rate constants for the reactions R1, R3, and R4, respectively and J 2 photolysis frequency of NO 2 .Previous studies investigating the NO-NO 2 cyclic system were conducted in urban and remote areas utilizing observations of O 3 (R1) and/or peroxy radicals (R3-R4) (Cantrell et al., 1997;Crawford et al., 1996 and references therein).They found that model-predicted peroxy radicals were often less than those required to explain the observed NO 2 /NO ratio.In addition, the model-predicted NO 2 levels were reported to be somewhat lower than observations.For example, comparison between observations and predictions ([NO 2 ] obs /[NO 2 ] calc ) in several field campaigns (CITE-3, ABLE-3B, CITE-2, and TRACE-A) typically showed 1.3 to 1.6 with a maximum of up to 3.4 in PEM-West A (Crawford et al., 1996).In contrast, Ridley et al. (1992) found good agreement between model peroxy radicals and those estimated from the NO 2 /NO ratio.Meanwhile, the potential role of iodine chemistry in NO 2 /NO ratio change as well as HO 2 /OH has been suggested based on field, model, and kinetic studies (Chameides and Davis, 1980;Davis et al., 1996;Knight and Crowley, 2001;Kanaya et al., 2002Kanaya et al., , 2007)).Since both hydroperoxyl and organic peroxy radicals were measured during the MIRAGE-Mex campaign, the photostationary state (PSS) of the NO-NO 2 system can be assessed based on field observations in this study without model estimates.
To assess the impact of MCMA emission on O 3 and its precursors on regional and hemispheric scales, it is important to evaluate the loss and transformation processes of reactive nitrogen species (NO x to NO y , which is total reactive nitrogen oxides) in the course of transport of urban or industrial plumes.The emitted NO x can be oxidized in the atmosphere by OH, forming HNO 3 which is subject to removal from the air mass through dry and wet deposition.The average lifetime of NO y in an urban, industrial (Nunnermacker et al., 2000) or continental outflow plumes (Takegawa et al., 2004) ranges from 0.25 to 2 days; whereas that of NO x in each environment is slightly shorter (less than 0.25 day in the former environment with longer lifetime in the latter environment, e.g., 1.2 days).
High levels of NO x in the urban plume result from enhanced local emission sources including transportation.Thus, NO x can be a measure of anthropogenic impacts at the sampling position and the ratio of NO x to NO y can be of value in understanding the chemical evolution process as an indicator for photochemical age (Carroll et al., 1992).In this study, we analyzed the NO-NO 2 cyclic system in different chemical regimes/air masses based on observations of reactive nitrogen species and peroxy radicals.The analysis of PSS allows us to assess the current understanding of tropospheric NO x chemistry including the potential for yet-unidentified chemical reaction.We also examined the NO x /NO y ratios and ozone production efficiencies of the polluted outflow from MCMA, providing some insight on the photochemical aging processes in various chemical regimes.

Observational data
The intensive field campaign of MIRAGE-Mex was carried out from 4 March to 29 March, involving the NSF/NCAR C-130 aircraft, ground-based measurements, and satellite observations.This campaign made numerous physico-chemical measurements such as reactive nitrogen species, oxidized sulfur species, oxygenated VOCs, aerosols, peroxy radicals, and so on.It also included the measurements of actinic flux and atmospheric photolysis frequencies such as J(O 3 ), J(NO 2 ), J(HNO 2 ), etc. (Shetter et al., 2002).In this study, we focused on the data of chemical measurements of reactive nitrogen species (NO, NO 2 , HNO 3 , PANs, organic nitrates, particulate nitrate, NO y ) and other trace gases such as O 3 , which were made on the C-130.Reactive nitrogen species, including NO, NO 2 , NO y , and O 3 were measured at 1 Hz with a chemiluminescence technique (Weinheimer et al., 1998).HNO 3 and organic nitrates were measured with a chemical ionization mass spectrometer (CIMS, Crounse et al., 2006 and references therein), while PANs were measured with a thermal dissociation CIMS (Slusher et al., 2004).As one of key measurements for the analysis of NO-NO 2 PSS (Eq.1), peroxy radicals (HO 2 +RO 2 ) were measured with the four-channel CIMS (Cantrell et al., 2003).Twelve C-130 missions were flown during the campaign, covering the altitude from the surface to about 7 km.Most of the flights sampled air over the MCMA basin and central Mexico, and several of the flights extended over the Gulf of Mexico to sample continental outflow from the MCMA.Table 1 shows detailed information on the 12 flights including spatial and temporal coverage.The data of flight 6 was excluded in the PSS analysis due to malfunction of the ozone instrument.More detailed flight tracks are available at the web site http://www.eol.ucar.edu/flightdata/mirage1/.those for FTCO and FTMA were based on geographical locations of airborne sampling.In addition, that for the BB was determined based on hydrogen cyanide (HCN), perchlorethene (C 2 Cl 4 ), and CO (Gregory et al., 1996;Li et al., 2000), while that for TIC was based on the concentration levels of CO, NO x , and SO 2 .The average concentrations of HCN, C 2 Cl 4 , and CO in the BB air masses corresponding to the flights on 4, 22, and 23 March were 965±761 pptv (median of 905), 4.2±7.5 pptv (0.92), and 278±184 ppbv (225), respectively.The median concentration of HCN for the BB is factors of 1.6 (TIC), 1.9 (BL), 2.5 (FTMA), and 2.8 (FTCO) higher than those for other air mass categories.The mean concentrations of SO 2 , CO, and NO x for the TIC corresponding to the data measured on 10, 19, and 22 March were 22±40 (median of 9.4), 186±81 (192), and 2.6±4.4 (0.82) ppbv, respectively.The mean concentration of SO 2 for TIC is higher than those for other air mass categories by factors ranging from 4.7 to 15.
The PSS of NO-NO 2 systems with different air mass categories was analyzed in Fig. 1.The PSS parameter (φ) is defined as follows: Printer-friendly Version Interactive Discussion In Eq. ( 2), the reaction rate constants (k 1 , k 3 , and k 4 ) are taken from Sander et al. (2002) and k 4 is the coefficient corresponding to the reaction between NO with CH 3 O 2 .If Reactions 1-4 represent adequately the NO-NO 2 partitioning, the value of φ (and the slope of Fig. 1) would be expected to be near unity.The PSS parameter (φ) ranged from 0.89 (FTMA) to 2.60 (BB) with the slope value of 1.13 for all data (Fig. 1a).The uncertainty (1σ) of the parameter is estimated to range from 23 (BL, FTCO, FTMA, and TIC) to 25% (BB) based on error propagation analysis using measurement uncertainties.In general, NO 2 /NO ratios are near PSS, showing a strong correlation (r 2 =0.73) with the value of k ).The discrepancy between the observed [NO 2 ]/[NO] ratios (denominator in Eq. 2) and calculated ratios (numerator in Eq. 2) was not always statistically significant for BL, FTMA, and FTCO.In contrast, there was slight deviation from the PSS for TIC.However, there was large deviation from the PSS for the BB air mass and the NO 2 /NO ratio in BB (5.18) was higher than those (2.35-4.17) in other air masses.The large deviation from the PSS for the BB might result from significantly less photochemical aging process (or relatively fresh plume), supported by lower ozone production efficiency (OPE of 4.6, detailed discussion in Sect.3.3) and higher NO x /NO y ratio (e.g., 0.3), compared to other air masses.More discussion on the ratio is given below.The discrepancy between the observed and calculated ratios can suggest the possibility of lack of current understanding of the tropospheric chemistry mechanism.The potential role of iodine chemistry in NO 2 /NO ratio change has been suggested based on field, model, and kinetic studies (Chameides and Davis, 1980;Davis et al., 1996;Knight and Crowley, 2001).Recently, significant impact of iodine chemistry (e.g., reaction of IO with NO producing NO 2 ) on the observed NO 2 /NO ratio has been reported for the field study at Rishiri island, Japan (Kanaya et al., 2002(Kanaya et al., , 2007)).The IO mixing ratio required to reproduce the PSS of NO-NO 2 system were estimated to be 0.8 pptv for FTMA on average.Similar ranges of IO levels (e.g., 0.3 at Cape Grim and 0.5-7 pptv at Mace Head) have been observed in the marine atmosphere (McFiggans, et al., 2000;Allan et al., 2000;Saiz-Lopez and Plane, 2004) and this suggests the potential role of halogen chemistry in NO 2 /NO ratio shift.
For the BB, O 3 and NO x concentrations were relatively higher with mean levels of 74 and 3.1 ppbv, respectively, higher than those for TIC, but less than those for the BL (79 and 3.7 ppbv, respectively) (Table 2).Mean concentrations of NO and NO 2 for the BB were 0.51 and 2.5 ppbv respectively.The mean NO 2 level for the BB was higher than those in other air mass categories by at least a factor of 1.3 and as much as a factor of 11 for the FTCO.The only exception to this was the BL.The relatively high NO x levels for the BB are likely to be affected by outflow from MCMA.NO x enrichment was also reported in the forest fire emissions near MCMA during the MIRAGE-Mex campaign on a different airborne platform (Twin Otter), ascribed to the deposition of nitrogen-containing pollutants in the outflow from the MC urban area (Yokelson et al., 2007).The concentrations of peroxy radicals, HO 2 and RO 2 for the BB were 58 and 159 pptv, respectively, which were factors of 1.3 to 4.7 (except for the BL) and 1.6 to 7.2 higher than those in other air mass categories, respectively.
It appears that there is no distinct trend in the PSS parameter φ in terms of altitude (above ground level and asl) (Fig. 2).In general, there was large variation in values of the parameter (φ) at altitude less than 3 km, as well as around 5 km for FTCO (Fig. 2d).
For FTCO, the mixing ratios of air pollutants such as O 3 and NO x were low in comparison to other air mass categories, with mean values of 51 and 0.3 ppbv.With increasing altitudes, the proportion of φ values less than 1 increases, possibly due to the increase of NO 2 /NO and/or the decrease of peroxy radical mixing ratios.Overall, there was no distinct PSS dependence on altitude.
According to a previous study of the PSS analysis of NO-NO 2 system (Crawford et al., 1996), the inverse value of the PSS parameter (φ improvements to the measurement technique during PEM-Tropics A (Bradshaw et al., 1999).Note that the number of data used in our PSS analysis was somewhat reduced due to lack of peroxy radical observations during the measurement period, compared to the NO 2 /NO and NO x /NO y data.
3.2 NO x partitioning for air mass category and photochemical aging The NO 2 /NO ratios for each air mass ranged from 2.35 (TIC) to 5.18 (BB), as shown in Fig. 3.There is significant correlation (r 2 =0.55 to 0.88) between NO 2 and NO in each air mass, indicating the same source in each air mass, supported by air mass back trajectory analysis (NOAA HYSPLIT model) (Fig. 4). Figure 4 shows only the most frequent air mass back trajectory for each air mass category.For the BL, westerly winds (W and SW) were predominant (75%) and southwesterly wind (75%) for the BB.
For the FTCO, westerly wind was dominant (35%) and southwesterly (41 and 67%) for the FTMA and TIC, respectively.The strong correlation (0.76-0.88) between NO 2 and NO was observed for the BL, BB, FTCO, and TIC air mass, whereas the FTMA category showed relatively weaker correlation (0.55).For the BL air mass, there was slight reduction in correlation intensity due to data on 26 and 29 March (data group in right-hand side (RHS) of curve fit, as shown in Fig. 3a).The re-calculated NO 2 /NO ratio (4.79 and r 2 =0.95) for the BL air mass increased when the data on 26 and 29 March were excluded.The air mass on 26 March for BL category is likely to be affected by Tula plume, supported by the somewhat similar NO 2 /NO ratio (1.54 and r 2 =0.99) to that (2.35 and r 2 =0.88) for the TIC and by air mass back trajectory analysis (downwind from TIC).In addition, the origin of the air mass on 29 March for BL seems to be changing from rural to urban.In other words, the air mass from about 12:00 to 13:00 local standard time (LST) (NO 2 /NO ratio of 1.81 and r 2 =0.97) originated from the rural air and from about 13:00-14:00 LST (ratio of 4.17 and r 2 =0.77) originated from the polluted MCMA air.The concentration level of NO x (2.3±4.3 ppbv with median of 0.94) in the data group of the RHS was similar to that for TIC.High levels of NO x and NO y within the megacity result from enhanced local emission sources such as combustion.Thus, NO x can be a measure of anthropogenic impact at the sampling position.In addition, the ratio of NO x to NO y can be valuable for understanding the chemical process as an indicator for photochemical age (Carroll et al., 1992), in spite of several limitations including the assumptions of no physicochemical loss (only dilution of NO y ), no chemical transformation from NO y to NO x (e.g., thermal decomposition of PAN), constant emission sources, and constant background concentration.In this field study, the NO x /NO y ratio varied from 0.17 (FTCO) to 0.54 (BL), depending on air mass category (Fig. 5).The higher ratios (0.38-0.54) for BL, BB, and TIC (Fig. 5a, b, and e, respectively) indicate less aged air than those (0.17-0.18) for FTCO and FTMA (Fig. 5c and d, respectively).The lower ratios for FTCO and FTMA indicated significant photochemical aging, which is supported by higher ozone production efficiencies, as described below.
The NO x /NO y ratios during MIRAGE-Mex were significantly higher than those (0.02-0.2) in continental outflow from East Asia during TRACE-P (Koike et al., 2003).However, the higher ratios (0.38-0.54) for BB, TIC, and BL were comparable to those in the near 2 day aged Asian outflow plume measured during the PEACE-A campaign (Takegawa et al., 2004).Our ratios (0.17-0.18) for the FTCO and FTMA were similar to the ratios (0.13-0.15) for marine air during PEM-Tropics B (Maloney et al., 2001).

Ozone Production Efficiency (OPE)
Photochemical oxidation of NO x and its oxidation products were intercorrelated with O 3 formation during the daytime.The OPE, which is the number of O 3 molecules produced per molecule of NO x oxidized (Kleinman et al., 2002 and references therein), was estimated using the linear regression slope between O 3 and oxidized NO x products (NO z =NO y -NO x ). Figure 6 shows the correlation between odd oxygen (O 3 +NO 2 ) and NO z and the OPEs for the 5 air mass categories in this field campaign ranged from 4.5 (TIC) to 8.5 (FTMA), indicating both photochemically young and reactive air masses.In general, higher efficiencies (5.9-8.5) were observed in FTCO and FTMA, 2284 Introduction

Conclusions References
Tables Figures Printer-friendly Version Interactive Discussion whereas lower OPEs (4.5-4.6) in TIC and BB.The OPEs in this field campaign were significantly lower than those (73-246) observed in the remote marine atmosphere such as the western North Pacific (Davis et al., 1996), but within the OPE ranges for the urban and rural atmosphere (Marion et al., 2001).

Summary and conclusions
Previous studies testing the NO-NO 2 cyclic system were conducted in urban and remote areas using observations of O 3 (R1) and/or peroxy radicals (R3-R4).These earlier studies used the concentrations of model-predicted peroxy radicals test the PSS for NO-NO 2 cyclic system and indicated that the peroxy radicals were not sufficient to explain the observed NO 2 /NO ratio.In addition, the model-predicted NO 2 levels in previous studies were reported to be somewhat lower than observations.In this study, we analyzed the NO-NO 2 system in different chemical regimes/air masses based on observations of reactive nitrogen species and peroxy radicals measured during the intensive field campaign of MIRAGE-Mex (4 to 29 March 2006).In general, NO 2 /NO ratios seem to be near PSS.For this analysis, the air mass was categorized into 5 groups such as BL, FTCO, FTMA, BB, and TIC.The NO 2 /NO ratios for each air mass ranged from 2.35 (TIC) to 5.18 (BB) and NO x /NO y ratio varied from 0.17 (FTCO) to 0.54 (BL), indicating both relatively photochemically young and reactive air masses (i.e., OPE ranges from 4.5 (TIC) to 8.5 (FTMA)).The potential role of halogen chemistry in NO 2 /NO ratio shift (i.e., lowering the PSS parameter, φ) can not be excluded according to our simple calculation of conversion of NO to NO 2 by IO in the marine atmosphere (FTMA case)., 37, 5207-5216, 2003. Takegawa, N., Kondo, Y., Koike, M., et al.: Removal of NO x and NO y in Asian outflow plumes: Aircraft measurements over the western Pacific in January 2002, J. Geophys.Res., 109, D23S04, doi:10.1029/2004JD004866,2004.Tie, X., Madronich, S., Li, G. H., Ying, Z., Zhang, R., Garcia, A. R., Lee-Taylor, J., and Liu, Y.: Characterizations of chemical oxidants in Mexico City: A regional chemical dynamical model (WRF-Chem) study, Atmos. Environ., 41, 1989-2008, 2007. Weinheimer, A. J., Montzka, D. D., Campos, T. L., et al.: −1 ) was reported to range from 1.33 to 3.36, estimated for several airborne sampling campaigns such as PEM-West A, CITE-3, ABLE-3B, CITE-2, and TRACE-A.From these campaigns, most φ −1 values were close to 1.4, but that for PEM-West A (3.36) was significantly higher in comparison.The cause for the significantly large deviation from the PSS in PEM-West A was suggested to be interference in the NO 2 measurement and this possibility stimulated

Table 2 .
A statistical summary of reactive nitrogen compounds, major air pollutants, and peroxy radicals with air mass categories measured during the MIRAGE fieldcampaign.