Tropospheric ozone sources and wave activity over Mexico City and Houston during MILAGRO / Intercontinental Transport Experiment ( INTEX-B ) Ozonesonde Network Study , 2006 ( IONS-06 )

During the INTEX-B (Intercontinental Chemical Transport Experiment)/ MILAGRO (Megacities Initiative: Local and Global Research Observations) experiments in March 2006 and the associated IONS-06 (INTEX Ozonesonde Network Study;http://croc.gsfc.nasa. gov/intexb/ions06.html ), regular ozonesonde launches were made over 15 North American sites. The soundings were strategically positioned to study inter-regional flows and meteorological interactions with a mixture of tropospheric O 3 sources: local pollution; O 3 associated with convection and lightning; stratosphere-troposphere exchange. The variability of tropospheric O3 over the Mexico City Basin (MCB; 19 N, 99 W) and Houston (30 ◦ N, 95 W) is reported here. MCB and Houston profiles displayed a double tropopause in most soundings and a subtropical tropopause layer with frequent wave disturbances, identified through O 3 laminae as gravity-wave induced. Ozonesondes launched over both cities in August and September 2006 (IONS-06, Phase 3) displayed a thicker tropospheric column O 3 (∼7 DU or 15– 20%) than in March 2006; nearly all of the increase was in the free troposphere. In spring and summer, O 3 laminar structure manifested mixed influences from the stratosphere, convective redistribution of O 3 and precursors, and O 3 from lightning NO. Stratospheric O 3 origins were present in 39% (MCB) and 60% (Houston) of the summer sondes. ComparCorrespondence to: A. M. Thompson (anne@meteo.psu.edu) ison of summer 2006 O 3 structure with summer 2004 sondes (IONS-04) over Houston showed 7% less tropospheric O 3 in 2006. This may reflect a sampling contrast, August to midSeptember 2006 instead of July-mid August 2004.


Introduction
A number of multi-national, multi-platform field experiments have been conducted to quantify intra-and intercontinental pollution transport.The 2004 ICARTT (International Consortium on Atmospheric Research on Transport and Transformation)/ INTEX (Inter-continental Transport Experiment -North America; Fehsenfeld et al., 2006;Singh et al., 2006) and 2006 MILAGRO/INTEX-B (Fast et  al., 2007; Molina et al., 2008 1 ; Singh et al., 2008 2 ) used multiple aircraft to follow pollutant transport into and out of North America.With the launch of the Aura spacecraft carrying four sensors in 2004 (http://aura.gsfc.nasa.gov;Schoeberl et al., 2007), tropospheric O 3 , a major pollutant, could be measured with some of its precursors (Zhang et al., 2006), providing a larger view of pollution transport.Although surface monitoring for pollutants like O 3 , CO and aerosol, is useful, trans-boundary and intercontinental transport often take place in the free troposphere (FT), just above the mixed layer (1-3 km) or in the mid-troposphere.The midtroposphere is also subject to direct injection of pollutants from convection, introduction of NO from lightning (Pickering et al., 1992;Ridley et al. 1996;Huntrieser et al., 2002) and stratospheric O 3 , frequently leading to thin layers and transport over long distances (Newell et al., 1999;Oltmans et al., 2004).Neither satellite nor aircraft instruments resolve pollutant structure from surface through the lower stratosphere, something that combined ozonesonde-radiosonde packages from balloons offer for O 3 .Thus, strategically designed sounding networks are an integral part of field campaigns.IONS-06 (INTEX Ozonesonde Network Study) operating during MILAGRO/INTEX-B (Phase 1=March 2006;Phase 2=mid-April-mid-May 2006;Phase 3=Augustmid-September 2006), supplied >700 O 3 and pressuretemperature-humidity (PTU) profile sets from surface to ∼35 km (5 hPa) at 23 North American stations (Table 1, cir-cles in Fig. 1).The IONS-06 soundings provided profiles for Aura instrument validation (e.g.Jiang et al., 2007;Schoeberl et al., 2007), for integrating field measurements and for evaluation of models (e.g.Pierce et al., 2007;Parrington et al., 2007).
The IONS-06 sonde network built on a similar effort, IONS-04, during the INTEX-A campaign; refer to the special INTEX-A/ICARTT (Singh et al., 2006;Fehsenfeld et al., 2006).IONS-04 was initiated to complement North American aircraft sampling and satellite measurements of O 3 , temperature and humidity (Thompson et al., 2007a, b), with eleven sites and sonde launches timed for Aura overpasses.IONS-04 configured sampling from the south central US through New England and the maritimes (Fig. 1, solid dots) to encompass major pollution sources and the route of eastward export from North America ( * symbol in Table 1 gives IONS-04 station coordinates).Combining forecasts with near-real-time O 3 data permitted INTEX-A and ICARTT aircraft to target sampling over North America, the Atlantic and Europe.Unexpected findings in IONS-04 were:

IONS-04/IONS-06 Locations
1. the prevalence of high-O 3 concentrations in the middle and upper troposphere (UT; Cooper et al., 2006;Thompson et al., 2007a) and 2. signatures of wave activity in stable laminae within most O 3 profiles (Thompson et al., 2007a, b).
The three-phase IONS-06 (sample distribution in In this paper, we quantify wave activity and tropospheric O 3 budgets over the MCB and Houston urban areas, using analytical approaches similar to the IONS-04 studies (Thompson et al., 2007a, b).With a focus on FT O 3 in Phase 1 (early spring) and Phase 3 (late summer), the following questions are addressed: 1. How does O 3 vertical structure over MCB and Houston vary day-to-day?Do we detect a similar mixture of influences (stratosphere, regional convection and lightning, advection) to those inferred in INTEX-A and IONS-04?
2. How do mean profiles and O 3 budgets between spring and summer compare?Are these consistent with meteorological patterns over Houston and the MCB, where the MILAGRO period has been summarized by Fast et al. (2007)?
3. What is the role of gravity waves, a prevalent dynamical signal, in determining upper troposphere and lower stratosphere O 3 over MCB and Houston?
4. How do the wave-affected laminae, tracers and trajectories translate into O 3 sources?Are there links in pollution between Houston and MCB?
For the MCB care is taken to use terrain following trajectories only.Additional meteorological data, satellite imagery and trajectory-based domain-filling products are taken from http://croc.gsfc.nasa.gov/intexb.Ozone and potential temperature (θ ) laminae, as described in Teitelbaum et al. (1994) and Pierce and Grant (1998) are used to identify signatures of Rossby (RW) or Gravity waves (GW).Wave frequencies at a given altitude for spring and summer are determined from the percentage of soundings within the season that have laminae with the RW or GW designation.The RW and GW classifications are refined in the "laminar identification (LID)" method to determine tropospheric O 3 budgets according to four contributions, as outlined in Thompson et al. (2007a): The use of an individual, dynamically derived BL height is an update to the Thompson et al. (2007a, b) method.Over MCB, the BL at mid-day launch averages ∼1.8 km above the surface.Fast et al. (2007) note that later in the day, under convective influence, the MCB BL may extend to 2.6 km above surface.Over Houston, the BL is ∼1 km, the value used in Thompson et al. (2007a, b).
A chemical "ozonopause" is used for LID calculations.In Thompson et al. (2007a) it was shown that although O 3 budgets and free tropospheric O 3 column amounts can differ significantly, depending on whether an ozone or thermal tropopause is employed, those occurrences were <10% in a typical set of the IONS soundings.Recently, a systematic comparison of ozonopause criteria (Dougherty, 2008) demonstrated how small differences are among five accepted methods, including the one used in the present paper.Over four summers' soundings over Houston, as well as other IONS-06 sites across North America, mean seasonal ST, RCL and AD budgets differed <5% absolute.

MCB and Houston ozone and relative humidity structure
Curtains of O 3 mixing ratio in 0.25-km averages over the MCB and Houston below 17 km (95 hPa), with the tropopause in white, appear in Fig. 2. The FT over MCB has moderate O 3 , 50-60 ppbv (Fig. 2a) with some clean layers, <35 ppbv, between 5 and 13 km.These layers originate mostly from tropical marine areas, with occasional recirculation over the nearby Caribbean (refer to trajectories at the IONS-06 website, http://croc.gsfc.nasa.gov/intexb/ions06).Marine origins are implicated by locally moist layers above 9 km in a number of the soundings.The FT over Houston (Fig. 2b) is sometimes as clean as over MCB but O 3 mixing ratios between 6 and 11 km average 10 ppbv higher over Houston than MCB (contrast Fig. 2c and d).Over MCB the tropopause is consistently high, averaging 15.5 km (Fig. 2a).
The O 3 structure between 13 and 16 km describes a subtropical tropopause transition layer with O 3 mixing ratio varying between 90 and 130 ppbv.More than half the March MCB and Houston profiles display a double ozonopause (see individual profiles at the IONS-06 website).For example, although Fig. 1 places a thermal tropopause at 15 km, Fig. 3a, for 11 March 2006, is suggestive of a double ozonopause, with the higher one at 125 hPa and the secondary one nearly coincident with the 250 hPa layer.On some days there are multiple laminae with stratospheric influence throughout the FT.For example, on 11 March 2006 two relatively dry layers over Houston (Fig. 3a), at 250 and 350 hPa, are distinguished by back-trajectories that pass over regions of elevated pv gradient over the Pacific (Fig. 3b).For the two days prior, the locations of several back trajectories, indicated by the "+" in Fig. 3b Mean O 3 , T and relative humidity (RH) profiles over MCB show that a moderately polluted mixed layer (to 5 km; Fig. 2c) gives way to a cleaner layer with relatively invariant O 3 up to 10 km (Fig. 2c).Above 10 km, the MCB O 3 mean and standard deviation increase.RH increases from 7 km to 10.5 km over MCB, suggesting surface origins, ie convective outflow in the mid-troposphere.A more complex tropopause structure over Houston, relative to MCB, displays a highly structured mean O 3 profile above 13 km (Fig. 2b,d Compare the recent study on double tropopause features in the sub-tropics (Randel et al., 2007).
the general pattern of higher O 3 over Houston above 11 km than over MCB (also drier over Houston, suggesting stratospheric influence), the mean T over Houston is greater than T over MCB.It is inferred that MCB has more tropical air with a colder tropopause layer.Indeed, back trajectories from the FT over MCB are consistently from the southwest, over the Pacific, whereas those over Houston often originate from the polluted Gulf Coast or western Texas.

Wave activity over MCB and Houston
Stable laminae associated with wave activity, as well as thin layers of advected pollution, are standard features in O 3 soundings.The morphology of wave types provides insight into processes affecting O 3 soundings.Higher correlation of θ-O 3 laminae, classified as GW by Teitelbaum et al. (1994) represents vertical displacements and potential convective activity.An RW designation indicates horizontal displacements, i.e. flow along isentropes, introducing stratospheric air into the troposphere.Figure 4 displays the frequency of GW and RW activity during March 2006 over the MCB and Houston.Combined GW and RW activity is most effective above 8 km for Houston (black lines), giving rise to more O 3 laminae and variability in the tropopause layer (Fig. 2d) than in the corresponding region over MCB (Fig. 2c).Over MCB, the increase of GW and RW in the subtropical tropopause layer is most pronounced above 12 km.The signature of RW activity above 12 km for MCB (Fig. 4) is consistent with the appearance of the very dry, high-O 3 layer at ∼160 hPa in virtually all the March soundings (see IONS-06 website).Over the MCB there is also a high GW frequency at 3-4 km, presumably where dry convection mixes BL and lower FT air.This is where Fast et al. (2007)  tive BL.Note two O 3 maxima over the MCB at 3.5 and 5 km (Fig. 2c), corresponding to the location of the GW feature (Fig. 4) and to the green shading, 50-60 ppbv, in the mixing ratio curtain (Fig. 2a).5c, d).Below 10 km over the MCB (Fig. 5c), RH increases greatly in August-September, the onset of the North American monsoon, compared to spring, the late dry season.For MCB the March 2006 RH averages 10-20% near the surface (Fig. 2c); the August-September mean exceeds 60% below 7 km (Fig. 5c).Compared to spring, mean O 3 mixing ratios at 8-12 km over both cities increase 20-35 ppbv.Over Houston (Fig. 5b) in August there is more yellow-orange-red above 6 km (>90 ppbv) as well as a more sharply defined tropopause (Fig. 5d), relative to spring (Fig. 2b, d).In summer, Houston O 3 profiles (August-September mean is the solid line in Fig. 5d) fall into a bimodal pattern.UT O 3 profiles with >100 ppbv features above 8 km are rare after 25 August 2006 (Fig. 5b).Pre-and post-25 August mean O 3 profiles (depicted as dashed, 1-25 August, and dotted, 26 August-11 September, with their respective RH profiles in Fig. 5d) have little overlap to within 1−σ (not shown).Cooper et al. (2006Cooper et al. ( , 2007)), compositing IONS-04 and IONS-06 profiles with MOZAIC (Measurements of Ozone by Airbus In-service Aricraft) landing/takeoff data from major airports (e.g.Dallas-Fort Worth, Atlanta), note that elevated UT O 3 appears to be a broad feature over summertime eastern North America.For the southern US IONS sites (Houston, Huntsville in 2004 and2006), Cooper et al. (2006Cooper et al. ( , 2007) ) use lightning flash data with FLEXPART to attribute most of the elevated O 3 in the 10-12 km layer to lightning.
For IONS-06, the southern US lightning interpretation (and significant UT O 3 from stratospheric sources over eastern North America), is further supported by analysis with a coupled chemical-transport model (Cooper et al., 2007;Hudman et al., 2007;Parrington et al., 2008).
Lightning-produced NO, along with post-convective redistribution of O 3 and/or O 3 precursors, has also been linked to FT O 3 layers over MCB in August-September 2006 (Ladino et al., 2007).This interpretation is consistent with a springto-summer RH increase above 12 km (deep convective outflow layers) over both cities (cf. Figs. 2c,d and 5c,d).Trajectory analysis (see IONS-06 website) shows that transport is even more decoupled between the MCB and Houston in   August-September 2006 than in March 2006.Most of the air parcels from MCB head west; origins of O 3 over MCB tend to be relatively localized.A few back-trajectories from Houston are from recirculation over the Gulf of Mexico; most are from the west.Flows away from Houston head north.

Spring 2006
The amount of tropospheric O 3 in layers associated with RW and GW combined is 40% and 44% over MCB and Houston, respectively, in March 2006.Similar values hold for August-September 2006.When filtered with tracers, the corresponding O 3 amounts translate to a combined ST and RCL fraction equivalent to 22% (MCB) and 32% (Houston) of the tropospheric column (Table 2).At Houston the ST fraction is twice that over MCB; convective influence (RCL in Table 2) is about the same over both cities.For the MCB the average tropospheric O 3 column is 11 DU lower than for Houston (Table 2), even though the 2-km thick BL at MCB has 1/3 more O 3 than the Houston BL.This appears to be due to a smaller FT O 3 column over MCB.
During March approximately half the O 3 is designated AD (Table 2), representing a mixture of recently imported O 3 pollution and a background amount of indeterminate origins, presumably mixed ST, RCL and earlier imported O 3 pollution.Figure 6

Summer 2006
Both the MCB and Houston increased in tropospheric O 3 column 7-8 DU in the spring-to-summer transition (Table 2).The BL O 3 amounts account for only 1 DU of the increase.Summer ST O 3 amounts (Table 2) did not change, although the ST fractions of total and free tropospheric O 3 decline 15-20% compared to the spring values.Stratospheric influence in the summer O 3 budgets (Fig. 6c, d) is significant, appearing in 39% of the MCB profiles and 60% of the Houston soundings.
Comparing  2).This is likely to be a consequence of a highly convective spring over Houston (Fast et al., 2007).The 500 hPa geopotential over southeastern Texas (not shown; refer to the website http://www.cdc.noaa.gov)show >50 m greater-than-normal heights in March 2006.The RCL and ST O 3 decreases from March to August-September 2006 over Houston are compensated for with higher AD, from 53% of the tropospheric O 3 column to 63%.All four O 3 component fractions change relatively little in the seasonal transition over MCB.Budget variations over Houston in August-September (Fig. 6c, d) reflect wave activity and the bimodal O 3 profile comparisons (Sect.4, Fig. 5c, d).Enhanced convection and lightning (higher O 3 concentrations and RH above 5 km) are more concentrated through 25 August (Fig. 5d, dashed profiles), corresponding to greater GW activity and higher RCL fractions (Fig. 6d).Lower midtropospheric O 3 mixing ratios (Fig. 5d, dotted profiles), a drier UT and a lower-altitude onset of the stratospheric O 3 gradient after 25 August 2006 coincide with more ST O 3 .
The pre-and post-25 August 2006 contrast in O 3 and RH Houston profiles (Fig. 5d) and in their corresponding LID budgets (Fig. 6d) is supported by meteorological analyses (Fig. 7).Prior to 25 August 2006, the fraction of RCL O 3 (14% of tropospheric column O 3 ) was more prevalent than the ST term (9% mean TCO fraction).The Houston region was dominated by an upper-level ridge as shown in Fig. 7a  Lower mid-tropospheric O 3 ratios, a drier UT (Fig. 6d) and a lower-altitude onset of the sharpest UT to lower stratosphere O 3 gradient after 25 August 2006 coincides with more ST O 3 (20%) than RCL O 3 (5%).A cut-off upper-level trough, depicted in Fig. 7c, moved slowly across the US, likely transporting O 3 from stratosphere to troposphere in early September 2006.
Ozonesondes over Houston were launched in IONS-04 as part of INTEX-A (July-August 2004; Morris et al., 2006;Thompson et al., 2007a, b).The IONS-04 budgets differ from IONS-06 (Table 2).The ST and RCL fractions of tropospheric O 3 in July-August 2004 were ∼25% higher than in 2006.Much of this difference is due to high RCL O 3 values from 2 August to 5 August 2004 when an anomalously high 500 hPa geopotential height and increased lightning exposure (see lightning exposure product, EL, at http: //croc.gsfc.nasa.gov/intexb)prevailed over east Texas.

Summary
Free tropospheric O 3 characteristics above the Mexico City Basin and Houston during spring and summer 2006 are delineated with sondes taken in IONS-06.Laminar analysis, tracer comparisons and meteorological interpretation explain the considerable day-to-day O 3 variability over both urban areas in both seasons.Wave influences are prominent in the laminar structure in the lower stratosphere and throughout the troposphere, with more frequent wave activity in spring.These signatures are consistent with vertical displacements by gravity waves (both sites) and mixing in the convective BL (Fast et al., 2007) over Mexico, where the UT/LS wave distribution resembles patterns observed over other subtropical sites (e.g.Grant et al., 1998;Loucks, 2007).
The soundings suggest that pollution transport between the Houston and Mexico City metropolitan areas during MILAGRO/INTEX-B (IONS-06 Phase 1, March 2006) was somewhat limited.This seems to agree with meteorological analyses (Fast et al., 2007) but may change with further studies of aircraft data (Molina et al., 2008 1 ; Singh et al., 2008 2 ).Trajectory analysis, combined with the soundings, showed even less coupling in the summer (IONS-06, Phase 3, TEX-AQS/GOMACCS) when both regions displayed a 7 DU increase in the tropospheric O 3 column (15-20% of the total).Most of the O 3 increase was registered in enhanced advection, based on our laminar identification technique, which means a higher background amount of O 3 or more likely, additional recirculation of regional pollution.
There is substantial stratospheric influence (ST) throughout the tropospheric O 3 profiles, over both sites and both seasons.In spring the relatively high ST is manifest in a double tropopause feature and a subtropical tropopause layer from 13-17 km.During summer the ST influences persisted, in 39% of the MCB soundings and 60% of those over Houston, where they were most intense after 25 August 2006.During 1-25 August 2006 over Houston, laminar analysis implicates a combination of convection/lightning and recirculating pollution in adding 20-30 ppbv O 3 to the troposphere above 8 km (cf.Cooper et al., 2006Cooper et al., , 2007;;Ladino et al., 2007).Moistening of the upper troposphere in early August also suggests a convective role, as does meteorological analysis.
After 25 August 2006 over Houston, convection subsides and a situation sets up that stimulates more ST influence.
The power of laminar identification to discriminate subtle but decisive impacts on O 3 structure, as in IONS-04 (Thompson et al., 2007a), is confirmed in this study.Along with complementary approaches, e.g.coupled chemical-transport models, trajectory-based analyses (e.g.FLEXPART), and aircraft tracers, a coherent picture of southern North American O 3 emerges.In this region, as in northeastern North America (Thompson et al., 2007b), the tropospheric O 3 structure is highly variable day-to-day.Free tropospheric O 3 in spring and summer 2006 is a rich mixture of ST and RCL influences.The persistence of ST O 3 influences in summer has also been noted in European sondes, based on trajectory and tracer analysis (Collette and Ancellet, 2005).

4
Figure 5a, b displays MCB and Houston O 3 mixing ratio curtains below 17 km in Phase 3 of IONS-06 along with mean O 3 , T and RH profiles (Fig.5c, d).Below 10 km over the MCB (Fig.5c), RH increases greatly in August-September, the onset of the North American monsoon, compared to spring, the late dry season.For MCB the March 2006 RH averages 10-20% near the surface (Fig.2c); the August-September mean exceeds 60% below 7 km (Fig.5c).Compared to spring, mean O 3 mixing ratios at 8-12 km over both cities increase 20-35 ppbv.Over Houston (Fig.5b) in August there is more yellow-orange-red above 6 km (>90 ppbv) as well as a more sharply defined tropopause (Fig.5d), relative to spring (Fig.2b, d).In summer, Houston O 3 profiles (August-September mean is the solid line in Fig.5d) fall into a bimodal pattern.UT O 3 profiles with >100 ppbv features above 8 km are rare after 25 August 2006 (Fig.5b).Pre-and post-25 August mean O 3 profiles (depicted as dashed, 1-25 August, and dotted, 26 August-11 September, with their respective RH profiles in Fig.5d) have little overlap to within 1−σ (not shown).Cooper et al. (2006Cooper et al. ( , 2007)), compositing IONS-04 and IONS-06 profiles with MOZAIC (Measurements of Ozone by Airbus In-service Aricraft) landing/takeoff data from major airports (e.g.Dallas-Fort Worth, Atlanta), note that elevated UT O 3 appears to be a broad feature over summertime eastern North America.For the southern US IONS sites(Houston, Huntsville in 2004 and 2006),Cooper et al. (2006Cooper et al. ( , 2007) ) use lightning flash data with FLEXPART to attribute Figure 5

Fig. 6 .
Figure 6 depicts tropospheric O 3 budgets computed by LID for daytime profiles over MCB and Houston in the March and August-September 2006 periods.In all cases there is considerable day-to-day variability in total tropospheric O 3 amounts as well as in individual budget fractions.The large and variable amounts of BL O 3 over MCB on 5-7 March and on 13 March 2006 (Fig. 6a) are partly due to a BL height >2 km above ground (not shown).Figure 6a indicates considerable ST O 3 over MCB on five days in March 2006.Over Houston ST O 3 is a consistent feature after 8 March 2006 (Fig. 6b).Individual GW fractions are ∼30% in early March 2006.After 8 March, a greater ST O 3 amount sets in, driven by a 20-40% RW frequency over the remainder of MILAGRO/INTEX-B.
Fig. 2,c and d with Fig. 5c, d reveals that most of the spring-to-summer O 3 increases occur between 5 and 12 km.Table 2 shows corresponding enhanced AD amounts, 5 DU over MCB and 8 DU over Houston.Recall that the AD term consists of recently imported or recirculated O 3 as well as an O 3 background that includes earlier ST, RCL and imported O 3 .Increases in relatively fresh imported or recirculated O 3 and an elevated background O 3 are both reasonable explanations for the higher AD O 3 .Between the March (MI-LAGRO, IONS-06 Phase 1) and August-September 2006 sampling periods, RCL and ST fractions over Houston decline ∼30% relative to their March contributions (Table , leading to frequent convection prior to 25 August 2006; an example, with clouds on 17 August 2006, appears in Fig. 7b.

Mexico City Basin Barbados R/V Ron Brown (IONS-06) Kelowna Stonyplain Bratt's Lake Egbert Walsingham Sable Is Yarmouth Boulder Richland Houston Socorro Paradox Table Mountain Narragansett Trinidad Head Valparaiso Wallops Is Figure 1 Pellston R/V Ron Brown (IONS-04) Holtville Huntsville Beltsville Fig
. 1. Map of IONS-04 (solid and starred) sites with the additional sites that comprised IONS-06 (open circles).See Table1for coordinates.The R/V Ronald H Brown ship operated in the Gulf of Maine in IONS-04 and in the Gulf of Mexico in Phase 3 of IONS-06 (August-September 2006).
(Butler et al., 2007)here the trajectory-mapped pv indicates stratospheric influence over several days.(InSect.5, it is shown that the LID method computes significant ST O 3 for the 9-11 March 2006 Houston profiles).The uv-DIAL O 3 instrument on the NASA DC-8 noted stratospheric air to 4 km over Houston on its 11 March 2006 MILAGRO/INTEX-B flight(Butler et al., 2007).Sondes and trajectories were used to explore the hypothesis that MCB and Houston O 3 pollution are linked (see profiles and trajectories on the IONS-06 website).Note that the sondes are somewhat limited compared to aircraft observations with multiple tracers.However, direct links between MCB and Houston are not easy to establish because comprehensive aircraft data coverage throughout the MILAGRO/ INTEX-B region is uneven.For the 11 March 2006 case noted above, there appears to be MCB influence in the Houston pollution O 3 layer at 750 hPa (corresponding trajectory in Fig.3b), but origins are actually north of MCB.The NASA DC-8 flight report for 11 March 2006 sampling over Houston, the Gulf of Mexico and MCB noted a complex O 3 struc-Trajectories forward from the MCB O 3 sounding on 16-18 March 2006, suggest MCB flows toward Houston and the northern Gulf of Mexico, but the layers were not notably polluted (see website).Forward trajectories from MCB on 17 March 2006 (700 hPa level, O 3 mixing ratio 50-60 ppbv, were predicted to pass over Houston 1-2 days later, but no Houston soundings were taken on 18 and 19 March 2006.The DC-8 detected moderate pollution influences from MCB near Houston and the Gulf (see http://catalog.eol.ucar.edu/milagro/report/dc-8/20060319/report.DC-8.200603191635.missionsummary.pdf).
Fast et al. (2007)98) influences in the larger region (http:// catalog.eol.ucar.edu/milagro/report/dc-8/20060311).An extended range of MCB impact was noted, based on uv-DIAL O 3 and aerosols(Butler et al., 2007), as well as other tracers.In some places biomass burning was implicated.In others, clean layers throughout the FT are noted.The complexity of flows and shifts in overall meteorology during MILAGRO are discussed byFast et al. (2007;see alsoFast and Zhong, 1998, for a larger context).The IONS-06 soundings encompass two periods of northerly flow, "Norte" events, discussed byFast et al. (2007); they ended prior to the third Norte episode.Rather than MCB air parcels heading toward Houston, during a Norte flows are in the opposite direction.
Table 2 displays mean budgets.