Chemistry and Transport of Pollution over the Gulf of Mexico and the Pacific: Spring 2006 INTEX-B Campaign Overview and First Results

. Intercontinental Chemical Transport Experiment-B (INTEX-B) was a major NASA (Acronyms are provided in Appendix A.) led multi-partner atmospheric ﬁeld campaign completed in the spring of 2006 ( http://cloud1.arc.nasa.gov/intex-b/). Its major objectives aimed at (i) investigating the extent and persistence of the outﬂow of pollution from Mexico; (ii) understanding transport and evolution of Asian pollution and implications for air quality and climate across western North America; and (iii) validating space-borne observations of tropospheric composition. INTEX-B was performed in two phases. In its ﬁrst phase (1–21 March), INTEX-B operated as part of the MILAGRO campaign with a focus on observations over Mexico and the Gulf of Mexico. In the second phase (17 April–15 May), the main INTEX-B focus was on trans-Paciﬁc Asian pollution transport. Multiple airborne platforms carrying state of the art chemistry and radiation payloads were ﬂown in concert with satellites and ground stations during the two phases of INTEX-B. Validation of Aura satellite instruments (TES, OMI, MLS, HIRDLS) was a key objective within INTEX-B. Satellite products along with meteorological and 3-D chemical transport model forecasts were integrated into the ﬂight planning process to allow targeted sampling of air parcels. Inter-comparisons were performed among and between aircraft payloads to quantify the accuracy of data and to create a uniﬁed data set. Pollution plumes were sampled over the Gulf of Mexico and the Paciﬁc several days after downwind transport from source regions. Signatures of Asian pollution were routinely detected by INTEX-B aircraft, providing a valuable data set on gas and aerosol composition to test models and evaluate pathways of pollution transport and their impact on air quality and climate. This overview provides details about campaign implementation and a context within which the present and future INTEX-B/MILAGRO publications can be understood.


Introduction
Air pollution is one of the most important environmental challenges of this century. This challenge is particularly acute in the developing world where the rapid growth of megacities (>10 million population) is producing atmospheric pollution of unprecedented severity and extent (Molina and Molina, 2004;Chan and Yao, 2008). In recent decades, a mounting body of atmospheric data have demonstrated that gas and aerosol emissions from major urban and industrial centers can influence both air quality and climate on scales ranging from regional to continental and global (Holloway et al., 2003;Stohl, 2004;Oltmans et al., 2006;EPA, 2007 and references there in). During the last decade, the NASA Tropospheric Chemistry Program has organized several field campaigns to investigate the outflow of pollution from Asia and North America with the goal of understanding the transport and transformation of gases and aerosols on transcontinental/intercontinental scales and their impact on air quality and climate (Hoell et al., 1996(Hoell et al., , 1997Raper et al.,  2001; Jacob et al., 2003;Singh et al., 2006). INTEX-NA was a major NASA led multi-partner field campaign that focused on the inflow and outflow of pollution over the North American continent. The first phase of INTEX-NA (INTEX-A) was completed in the summer of 2004 under the ICARTT umbrella and results widely disseminated Fehsenfeld et al., 2006 (Hoell et al., 1997;Jacob et al., 2003;Parrish et al., 2004;Molina et al., 2007;Nowak et al., 2004). INTEX-B/MILAGRO was an integrated mission performed principally over Mexico, the Gulf of Mexico, and the northern Pacific Ocean during the spring of 2006 in cooperation with multiple national and international partners (Molina et al., 2008). Major objectives for INTEX-B were aimed at (i) investigating the extent and persistence of the outflow of pollution from Mexico; (ii) understanding the transport and evolution of Asian pollution and implications for air quality and climate across western North America; (iii) relating atmospheric composition to sources and sinks; (iv) characterizing the effects of aerosols on solar radiation; and (v) validating space-borne observations of tropospheric composition. This effort had a broad scope to investigate the chemistry and transport of long-lived greenhouse gases, oxidants and their precursors, aerosols and their precursors, as well their relationship with radiation and climate.

Measurement platforms
The principal mobile platforms that participated in the INTEX-B/MILAGRO campaign, their bases of operation, campaign duration, and specifications are summarized in Table 1. These airborne platforms had the capability to sample much of the troposphere over long distances. Complementing these airborne platforms were several ground stations that operated during this study. An ozonesonde-radiosonde network (IONS-06) operated over 16 North American sites. Satellite observations both guided flight planning and helped to extend the range of airborne observations. Description of Mexico based platforms and ground stations focused exclusively on the MI-LAGRO part of the campaign are being summarized in an overview paper by Molina, L. and Madronich, S., private communication, 2009. Additional details about these platforms are available from http://cloud1.arc.nasa.gov/intex-b/ and http://www.eol.ucar.edu/projects/milagro.

Instrument payload and measurement capability
The payloads of a majority of airborne platforms in Table 1 were selected to measure major greenhouse gases, ozone and precursors, aerosols and precursors, and a large number of tracers. Table 2a shows the DC-8 payload of in situ and remote sensors for measurements of both gases and aerosols. A nadir and zenith viewing UV lidar measured tropospheric O 3 and aerosols remotely. A second zenith viewing lidar was tuned to stratospheric ozone and temperature measurements for HIRDLS validation. Spectral radiometers allowed direct   pre-concentration on Au trap w/UV fluorescence 0.1 ng/m 3 10-15% (Tekran, Inc.) Standard met and GPS temperature, RH, wind speed, wind direction sensors a These are nominal values. Detailed measurement specifications, which may vary with altitude, are provided in the headers along with data.   (Table 3). These launches were coordinated with the daytime Aura overpass .

Flight planning and execution
Satellite observations, along with meteorological and chemical forecasts from global and regional chemical transport models (CTMs), were extensively used to plan, coordinate, and implement the INTEX-B field mission. Table 4 lists the major CTMs in use for planning this experiment. Meteorological products were derived from the NCEP reanalysis data. Daily meteorological and chemical forecasts guided the design and execution of the mission. Typical model products included source-tagged CO tracers, aerosol tracers, O 3 , and PAN distributions. Meteorological products included satellite imagery (visible, infrared, and water channels), precipitation and cloud fields, backward and forward trajectories, and multi-day synoptic weather forecasts. Principal satellite instruments that provided near real-time data to guide flight planning were AIRS and MOPITT (CO columns), MODIS (aerosol, fire counts), and OMI (total column O 3 , tropospheric NO 2 , absorbing aerosols). These data were combined with forward trajectories for flight planning purposes.
The NSF/NCAR C-130 coordinated activities with the DC-8 with bases in Mexico and Seattle. This coordination was necessary to accomplish inter-comparisons among instruments, to achieve several objectives that were unattainable by a single platform, and to collect complementary data where appropriate. This also permitted quasi-Lagrangian investigations when pollution plumes sampled over Mexico and the Pacific were sampled again after several days of transport.  Table 5a and b provides a brief summary of the salient flight objectives of the DC-8 and C-130 with more details available at http://cloud1.arc.nasa.gov/intex-b/. Air masses influenced by anthropogenic pollution, forest fires, and stratosphere were sampled during INTEX-B. It was possible to fly extremely low (100 m a.g.l.) for source characterization and spiral up to 12 km for purposes of satellite validation. Many of the flight planning procedures used in INTEX-B were similar to INTEX-A and have been previously described .

Inter-comparisons
Intra-and inter-platform comparisons of measurements were extensively performed to establish accuracies and precisions. On the DC-8 itself, several chemical measurements were duplicated using a variety of different techniques. Salient among these were measurements of O 3 , H 2 O, HCHO, and jNO 2 . Inter-comparisons were also performed with other aircraft by flying in formation with a separation of less than 300 meters in the horizontal and 100 m in the vertical. Two to three levels were chosen between 0 and 6 km altitudes, and the entire inter-comparison including descent and ascent typically lasted one hour. Table 6 provides a summary of the inter-comparisons carried out by the airborne platforms, which included species such as O 3 , OH, HO 2 , CO, SO 2 , PAN, NO 2 , HNO 3 , NO y , VOC, and aerosol composition and properties. Comparisons were conducted in two phases: a blind phase occurring in the field shortly after completion of the flight, and a final phase after data was fully calibrated and approved by the investigators. The actual inter-comparison data and results are available at http://www-air.larc.nasa.gov/ cgi-bin/ic. While the agreement among instruments on various platforms was generally good for molecules such as O 3 , NO 2 , CO, H 2 O, CO 2 , HCHO, and aerosol physical properties, there were large differences in measurements of HO x , OVOC, HNO 3 , PAN, NO, and H 2 O 2 that suggest a need for concerted efforts to reduce these errors.

Satellite validation
Satellite sensors measure radiances from which information on atmospheric composition is retrieved using radiative transfer algorithms with substantial and often poorly characterized uncertainties. Validation of satellite derived tropospheric composition and optical properties were a pri-mary goal and received a great deal of attention from the NASA DC-8, J-31, and to a lesser extent, the C-130 platforms. The principal focus of INTEX-B was on the Aura satellite (Schoeberl et al., 2006), but validation studies for SCIAMACHY/Envisat, MODIS-Terra/Aqua, and MISR/Terra were also carried out (Table 7a, b). The DC-8 focused on several trace gases (O 3 , CO, NO 2 , HNO 3 , and HCHO) and together with the J-31 on aerosols and clouds.
Opportunities to under-fly satellites for integrated science and validation were examined and planned for each flight.    Validations were performed to test retrievals with a variety of underlying surfaces. A typical DC-8 validation flight for nadir viewing instruments (e.g. OMI, TES) involved a profile from 0.2 to 11.5 km, spiraling with a 20 km radius within ±30 min of the satellite overpass time under relatively cloud free conditions. Limb viewing instruments (e.g. MLS) frequently required sampling along the satellite track utilizing both in situ and remote sensing capabilities on the DC-8. Two flights were dedicated to validation of HIRDLS O 3 and temperature in the stratosphere using the AROTAL lidar, which required night flights for observations to reach maximum altitudes. In all cases, the DC-8 provided important chemical and physical data that could be used to improve satellite retrievals. While satellite under-flights were usually planned to focus on a specific satellite instrument, the prox-imity of A-Train observations and wide swaths for some instruments allowed these flights to routinely benefit multiple satellite sensors.

Overview of first results
We provide here a brief overview of the INTEX-B results presented in the first collection of papers assembled in this Special Issue of Atmospheric Chemistry and Physics and published elsewhere to date. We focus principally on the Pacific component assessing the influence of transpacific transport. Additional information on the INTEX-B/MILAGRO mission is available from Molina et al. (2008) and a complementary overview paper that focuses on activities from platforms based within Mexico is in preparation (Molina, L. and  Madronich, S., private communication, 2009). In the following sections we briefly discuss results based on (1) integrated analysis of data using observations and models; (2) chemistry and transport model evaluation; and (3) satellite validation results and analysis from satellite observations. The reader is referred to the specific papers for more information. We also note that the data continue to be further analyzed and new results are anticipated.

Pollution transport and chemistry: integrated analysis of data
Many studies being published in the ACP Special Issue analyzed observed data from aircraft, ground stations, and satellites with the help of models. Here we briefly discuss

Airborne observations and trans-pacific transport
There was extensive evidence for pollution transport from Asia to North America and all the way to Europe. Concurrent satellite CO and ozone observations (TES, AIRS, MOPITT) showed direct evidence for trans-Pacific pollution transport (Molina et al., 2008;Zhang et al., 2008 INTEX-B data over the Pacific clearly showed that a dominant fraction of reactive nitrogen was in the form of transported PAN. Wolfe et al. (2007) utilize observations of PANs at the MBO to conclude that trans-Pacific transport of Asian pollution leads to substantial increases in PANs and ozone at this site. The ensemble of trajectories indicates that Asian-influenced free tropospheric air contained a median PAN mixing ratio double that of the full dataset.  use observations and 3-D models to elucidate the important role PAN plays in O 3 formation resulting from its ability to redistribute NO x . Measurements of CO at MBO and meteorological indices show that long-range transport of CO from the heavily industrialized region of East Asia was significantly lower in 2006 compared to 2005 . In the Western US both ground site data and MO-PITT and TES satellite observations reveal a significant decrease (from 2-21%) in springtime maximum CO between 2005 and 2006. They attribute this decrease to moderate biomass burning in Southeast Asia during 2006 and a transport pattern that limited the inflow of Asian pollution to the lower free troposphere over western North America.
Analysis of aircraft sulfate measurements from the DC-8 over the central Pacific, C-130 over the east Pacific, and the Cessna over British Columbia indicated that most Asian aerosol over the ocean was in the lower free troposphere (2-5 km). Large amounts of sulfate and little organic carbon were found in imported Asian pollution over Western Canada (Fig. 3). Asian plumes were not only significantly reduced of fine particle organic material, but organic compounds were attached to coarse particles in significant quantities. Scavenging of organic aerosol precursors by dust near source regions is suggested, and any formation of secondary organic aerosol during transport from Asian source regions across the Pacific was principally associated with the coarse particles McKendry et al., 2008). In agreement with Zhang et al. (2008), an average of profiles indicated that trans-Pacific transport between 2 and 5 km during this period increased ozone by about 10 ppb and fine particle sulfate by 0.2-0.5 µg m −3 . Campaign-average simulations showed that anthropogenic East Asian sulfur emissions increased mean springtime sulfate in Western Canada at the surface by 25-30% and accounted for 40% of the overall regional sulfate burden between 1 and 5 km . Mc-Naughton et al. (2009) analyze data from the eastern Pacific and deduce that the presence of CaCO 3 in Asian dust enhanced heterogeneous conversion of gas phase S and N into aerosol phase.
Fine particle median sulfate and water-soluble organic carbon (WSOC) concentrations in transported Asian air masses (2-5 km) were two to four times lower than North American air masses. In contrast, in air masses below 2 km median WSOC-sulfate ratios were consistently between one and two . Sun et al. (2008) use measured aerosol mass spectra at Whistler peak to conclude that sulfate and OA were mostly present in external mixtures, indicating different origins, and OA was highly oxygenated, with an average organic-mass-to-organic-carbon ratio of 2.0 and an atomic ratio of oxygen-to-carbon of 0.8. The nominal formula for OA was C 1 H 1.52 N 0.03 O 0.82 for the entire study. Heald et al. (2008) synthesize measurements of organic carbon compounds in both the gas and particle phases made upwind, over, and downwind of North America. The daytime mean total organic carbon ranges from 4 to 456 µgC m −3 with organic aerosol making up 3-17% of this mean. Anthropogenic emissions account for much of this variability but correlation with tracers such as isoprene suggest that biogenic emissions are an important contributor.  describe two case studies for pollution layers transported across the Pacific from the Asian continent, intercepted 3-4 days and 7-10 days downwind of Asia, respectively. These observations of sulfate and organic aerosol in aged Asian pollution layers are consistent with fast formation near the Asian continent, followed by washout during lofting and subsequent transformation during transport across the Pacific. No evidence is found to support significant secondary organic aerosol formation in the free troposphere.
A number of specialized measurements, new to the DC-8, were performed to define Pacific composition for the first time. Kim et al. (2008) provide specific gas phase HCl measurements from the marine boundary layer (MBL) to the lower stratosphere. They find that HCl mixing ratios are lower than previously believed and show evidence for its production in the mid troposphere by the de-chlorination of dust aerosols. Over the Pacific, mercury was weakly correlated with anthropogenic tracers . A prominent feature of the INTEX-B dataset was frequent total depletion of Hg • in the upper troposphere when stratospherically influenced air was encountered (Radke et al., 2007;Talbot et al., 2007;. Global 3-D chemical transport models have been used to interpret new INTEX-B observations of methanol to further constrain the atmospheric methanol budget . INTEX-B provided a most detailed exploration of HO x free radicals and its precursors. Simulation studies highlight many uncertainties in our quantitative understanding of the HO x -NO x -O 3 system. Mao et al. (2009) deployed a new airborne OH reactivity instrument for the first time on the NASA DC-8 during the INTEX-B campaign. From the median vertical profile obtained over the Pacific, the measured OH reactivity was 30% higher than the OH reactivity from steady state assumptions and some 200% higher than the OH reactivity calculated from the total measurements of all OH reactants. They attribute the missing OH reactivity to some highly reactive unmeasured VOCs that have HCHO as an oxidation product.

Model simulations and INTEX-B observations
INTEX-B provided a vast amount of observational data to both test CTMs and use them for further analysis. In general, the models in Table 4 captured the presence of pollution plumes and were an excellent aid in flight planning. However, significant quantitative differences remain. Figure 4a shows a comparison of selected observed constituents (OH, CO, and O 3 ) and results from post mission simulations over the Gulf of Mexico (Region 1) and Pacific (Region 4). As was previously observed in INTEX-A (Singh et al., 2007), model disagreements far exceed any potential error  Boersma et al. (2008), Luo et al. (2007) and Nardi et al. (2008). See text for more details.
in observations and are also not consistent between missions (INTEX-A and INTEX-B). In case of GEOS-Chem, overestimation of OH appears to be a feature that is propagated throughout the INTEX-B Gulf and Pacific domain, and is principally responsible for CO underestimation. The "box model", which is most constrained by observations, is consistent with measurements above ∼4 km but significantly over predicts OH in the lower troposphere. Disagreements with observations are often much larger for more complex organic molecules such as acetaldehyde (Fig. 4b). Comparisons with CTMs of data from the entire campaign reveal an underprediction of organic aerosol mass in the MOZART model, but much smaller discrepancies with the GEOS-Chem model than found in previous studies over the western Pacific . Model disagreements can be due to a large number of factors that include emissions, boundary conditions, chemistry, meteorology, and even computational errors that need to be fully investigated. Plans are presently underway to create a community model inter-comparison and test project within the IGAC umbrella.

Validation results
Providing coincident observations for satellite validation of Aura instruments was a central theme for the DC-8 in both phases of INTEX-B. A limited number of validation comparisons are summarized in Fig. 5 and more details referenced. Figure 5a, b shows a correlation plot of MLS re-trieved CO and O 3 in the UT (215 hPa) with the DC-8 data indicating large differences in CO and relatively good agreement with O 3 . Figure 5c compares tropospheric NO 2 column measurements from OMI with airborne observations over the Southern United States, Mexico, and the Gulf of Mexico during the INTEX-B campaign. Reasonably good correlation with no significant bias (R 2 =0.67, slope=0.99) is found for the ensemble of comparisons when the aircraft could spiral sufficiently low to sample most of the NO 2 column Bucsela et al., 2008). TES CO data show good (±20%) overall agreement in the troposphere with the DC-8 in situ measurements ( Fig. 5d) with some biases in the UT and LT (Luo et al., 2007). Comparisons of the TES O 3 with the DC-8 DIAL (Fig. 5e) and ozonesonde data show a somewhat positive bias for TES (Richards et al., 2008;Nassar et al., 2008). Santee et al. (2007) provide an in depth comparison of limb derived MLS HNO 3 data with that of INTEX-B and other observations.
Although INTEX-B primarily focused on the troposphere, one of the DC-8 remote sensing instruments (AROTAL) was targeted for stratospheric observations. Figure 5f shows a comparison of the DC-8 AROTAL and HIRDLS retrieved O 3 with good agreement above 20 km and poor agreement in the lower stratosphere. Nardi et al. (2008) provide details on these comparisons and the reasons for the disagreements. Froidevaux et al. (2008) investigate MLS O 3 retrievals in the stratosphere and compare them with the INTEX-B AROTAL data.  Fig. 2a. Terrain effects in Region 2 cause the column to be 2-12 km above surface. Missing data in column calculation were interpolated. * * Mean ± one sigma.
Extensive validation of CO from AIRS and MOPITT was also carried out. Emmons et al. (2008) use INTEX-B/MILAGRO and other in situ observations to conclude that a significant positive bias, around 20%, is present in MO-PITT derived CO. Comparisons to the long-term records revealed that this bias is increasing with time. Possible causes for this drift and improvements in future retrievals have been suggested. Felker et al. (2008) blend GOES water vapor and TES ozone to derive a Multi-sensor UT Ozone Product (MUTOP) that has the advantage of producing the spatial coverage of a geostationary image while retaining TES's ability to resolve UT ozone. Results show 70-80% of TESobserved UT ozone variability can be explained by correlation with these two dynamical tracers. It is noted that satellite retrievals are continually being refined and it is possible that the comparisons shown here will improve over time.

Satellite based analysis
OMI NO 2 satellite observations were used to constrain Asian anthropogenic NO x emissions, indicating a factor of two increases in emissions from China from 2000 to 2006 . Halland et al. (2008) quantify the vertical transport of CO by deep mesoscale convective systems and assess the ability of TES to detect the resulting enhanced CO in the upper atmosphere. A squall line similar to one in INTEX-B was simulated using the Goddard Cumulus Ensemble model. Results show that the simulated squall line transports the greatest mass of CO in the UT. They conclude that TES has sufficient sensitivity to resolve convectively lofted CO, as long as the retrieval scene is cloud-free. Arellano et al. (2007) present a global chemical data assimilation system and apply it to constrain global tropospheric CO by assimilating meteorological observations of temperature and horizontal wind velocity and satellite CO retrieved from MOPITT. They conclude that MOPITT data assimilation provides significant improvements in capturing the observed CO variability. Shim et al. (2009) use concurrent tropospheric O 3 and CO vertical profiles from TES over Mexico City to characterize mega-city pollution outflow on a regional scale. They find that TES captures much of the spatial and day-to-day variability seen in the in situ data for O 3 , but not for CO and cautions against using such correlations. Y. Shinozuka and co-workers (private communication, 2008) report that number concentration of particles >100 nm are a proxy for CCN and can be stratified by two optical properties, the extinction coefficient and its wavelength dependence. They use data from INTEX-B/MILAGRO and INTEX-A to quantify these relationships and explore the feasibility of satellite retrieval of CCN concentration from optical property measurements.
It is increasingly possible to retrieve many organic and inorganic species from existing satellite obtained spectra (e.g. acetone, methanol, HCHO, PAN, CO 2 , CH 4 ) (Beer et al., 2001;Frankenberg et al., 2005;Glatthor et al., 2007;Rinsland et al., 2007). In Table 8, we provide UT mixing ratios and tropospheric columns of key atmospheric constituents that are potentially retrievable from existing satellite data for the four selected regions of INTEX-B (Fig. 2a). The variability in the CO 2 column and UT mixing ratios is 2.7% (≈10 ppm) and 1.5% (≈6 ppm), respectively, irrespective of the nature of underlying sources. Filtering the data for only clean conditions (bottom two CO quartiles) had little effect (<5%) on the relative variability. Boundary layer (1-2 km) CO 2 elevation due to urban pollution and reductions due to biosphere uptake would need to be 40 ppm or more to be detected in the CO 2 column. Airborne observations indicate that these elevations are typically less than 15 ppm and only on rare occasions may exceed 40 ppm (Choi et al., 2008). Relating satellite derived atmospheric columns of CO 2 (Orbiting Carbon Observatory; http://oco.jpl.nasa.gov/; Miller et al., 2007) to its surface sources and sinks would be extremely challenging.

Conclusions
The INTEX-B/MILAGRO measurements provide a rich data set for describing the composition and chemistry of the troposphere over Mexico, the Gulf of Mexico, the Pacific, and western North America, for investigating the transformation of gases and aerosols during long-range transport, for the radiation balance of the troposphere, and for validating a variety of satellite observations as well as models of chemistry and transport. The results presented in this special issue of ACP represent only the initial analysis, and much more integrated analysis is expected in the near future. The data are available to all interested parties for further analysis.