Nonmethane hydrocarbon and halocarbon distributions during Atlantic Stratocumulus Transition Experiment/Marine and Gas Exchange,

. Aircraft measurements of selected nonmethane hydrocarbon and halocarbon species were made in the lower troposphere of the NE Atlantic near the Azores, Portugal, during June 1992 as part of the Atlantic Stratocumulus Transition Expefimen(cid:127)arine Aerosol and Gas Exchange. In this paper, the impact of continental outflow from both Europe and North America on the study region were assessed. Four main air mass types were characterized from trajectories and trace gas concentrations: clean marine from the Atlantic, and continental air from the Iberian Peninsula, the British Isles and Northern Europe, and North America. Each classification exhibited trace gas concentrations that had been modified en route by photochemical processes and mixing. Comparison with the clean marine boundary layer (MBL) shows that the boundary layer of the predominantly continental air masses were enhanced in hydrocarbons and halocarbons by factors of approximately 2 for ethane, 5 for propane, 2-6 for ethyne and benzene, and 2-3 for C2C14. The same air masses also exhibited large ozone enhancements, with 2 to 3 times higher mixing ratios in the continental boundary layer air compared to the clean MBL. This indicates a primarily anthropogenic photochemical source for a significant fraction of the lower tropospheric ozone in this region. Methyl bromide exhibited on average 10-20% higher concentrations in the boundary layer affected by continental outflow than in the clean MBL, and was seen to be enhanced in individual plumes of air of continental origin. This is consistent with significant anthropogenic sources for methyl bromide. In addition, median MBL concentrations of ethene and methyl iodide showed enhancements of approximately a factor of 2 above free tropospheric values, suggesting primarily coastal/oceanic sources for these species.


Introduction
Nonmethane hydrocarbons (NMHCs)and halocarbons can play critical roles in tropospheric chemistry, and many of the halocarbons are important as infrared absorbers of terrestrial radiation and are halogen sources to the stratosphere. Anthropogenically produced halocarbons are released predominantly in urban areas and may be used to estimate the recent anthropogenic influence of an air parcel [Blake et al., 1992[Blake et al., , 1994Smith, 1993].
The concentrations of selected chlorofluorocarbons (CFCs) and other halocarbons have been published for Ireland [Prinn et al., 1983;Simmonds et al., 1993] and for the Pacific [Elkins et al., 1993;Wang, 1993]. The variation of tetrachloroethene in the Pacific has been characterized by Wang et al. [1995], and in the North Atlantic for tetrachloroethene and methyl chloride by Koppmann et al. troposphere. Methyl iodide has been suggested as the main carrier of iodine from the ocean to the atmosphere [Lovelock et al., 1973].  were evacuated to a pressure of 10 '2 torr and shipped via air cargo to Santa Maria Island. After conclusion of the flights, the canisters were sent for analysis to a temporary laboratory established by our research group at the NOAA Atlantic Oceanographic Meteorological Laboratory in Key Biscayne, Florida. The location of this laboratory was dictated by the logistical requirements of the NASA TRansport and Chemistry near the Equator (TRACE-A) project which came immediately after the conclusion of ASTEX/MAGE. The laboratory was in operation from July through November 1992. Outside air was brought into the NCAR Electra through a 1/4inch stainless steel inlet mounted on the port side of the fuselage just before the wing, and extended out 12 inches, sufficient to reach beyond the laminar boundary layer of the aircraft. This inlet was attached in series to a mass flow controller (Brooks Instrument Division, model 5850 E series), a two-stage metal bellows pump (Metal Bellows Co., MB-602), and a gas-handling manifold. The manifold was mounted on the base of the rack and directed the air flow to each of the three 24-canister sets. Air was continuously pumped through the manifold to flush the inlet lines and canister tubing before exiting the aircraft through an outlet mounted directly behind the inlet. This configuration allowed the system to be operated from subambient pressures to 45 psig, above which a pressure relief valve was activated. Depending on the mass flow controller setting and the ambient pressure associated with a particular altitude, the time required to pressurize a sample to 40 psig ranged from 10 to 120 s. Therefore, depending on the speed and the ascent/descent rate of the Electra, each sample represents an air mass sampled over a horizontal distance of 0.7 to 10 km or a vertical distance of up to 300 m.

Chemical Analysis
Each whole air sample was analyzed for NMHCs and halocarbons using a trace gas apparatus with three gas chromatographic separation columns contained in two temperature programmed Hewlett-Packard 5890 Series II gas chromatographs. Four separate gas chromatograms were obtained for each sample. To monitor the consistency and reproducibility of the analytical system, a secondary working standard contained in an Aculife treated Luxfer cylinder (dry ambient air collected at Niwot Ridge, Colorado) was analyzed after every four samples. To ensure there was no drift in this working standard, two other whole air cylinders were assayed randomly throughout the project. Concentrations of reported gases in the working standard exhibited no statistically significant changes (1 sigma) in their mixing ratios during 379 measurements over the course of the project.
Became the concentrations for some of the gases in the working standard were significantly higher than those routinely encountered during the missions, our measurement precision for the gases assayed during ASTEX are based on 4 samples collected in the MBL during a 16 minute period of mission 11. The mean mixing ratios, along with the corresponding 1-sigma standard deviations were (in pptv) ethane 1112+6; ethene 43.8+5.9; ethyne 298.3+4.7; propane !44.0d:1.9; propene 20.2+2.1; nbutane 69.1+1.9; i-butane 53.5+1.6; n-pentane 11.0+1.0; benzene 103.5ñ3.2; and toluene 41.7+2.9. Propyne and nhexane were below their 3 pptv detection limit. The mean halocarbon mixing ratios and standard deviations for these samples were ( Table 1.

Most flights contained multiple altitudinal profiles to aid in the characterization of chemical variability of the vertical column. Individual flight information, including sampling periods, predominant origins of air masses from trajectory information, approximate sampling areas, and brief mission descriptions are given in
Meteorological descriptions of ASTEX/MAGE can be found in the work of Bretherton and Pincus [1995]. In summary, the flights were conducted near the North Atlantic surface highpressure system, the center of which was located to the west of the sampling region. This area was associated with very clean maritime air that was sampled during missions 5, 6 and 7. During the early part of the deployment, transport around this system brought modified (i.e., diluted and photochemically aged) continental air from North America during missions 1, 3, and 4.

Mission 2 was conducted nearer the center of the surface high and was mainly influenced by clean marine air. Later (missions 8 to 10), a surface low moving off the coast of Portugal brought
European air from the direction of the Iberian Peninsula to the project area. A ridge associhted with a high-pressure system then strengthened to produce northeast winds that transported continental air mainly from the British Isles to the area of operation for missions 11-15. Table 2 shows     Box plots comparing the lVlI3L concentration ranges for these two types of Continental European air masses are given in Figure  5, along with the ranges for two further air mass types described later. Figure 5 shows that the most noticeable difference in the chemical composition of the two European air masses is the significantly higher concentration of CH3CCI 3 found in the northern case (missions 11-14). By contrast, the median concentrations of C2C14 and most of the hydrocarbons are slightly lower in the air from northern Europe. Methyl chloroform was widely used as a general purpose solvent, so the higher concentrations that appear to originate from this region are probably the consequence of different types of industrial use.
The   7. They were made in clean Ariantic air associated with the strong high-pressure system near the Azores, and the air mass is labeled as clean marine in Table 1 and Figure 5. In general, the high-pressure system circulates air over the ocean, where it becomes well mixed and does not come into contact with land for many days. Consistent with this, the MBL was found to be chemically homogeneous. The shorter-lived anthropogenic species n-butane, i-butane and n-pentane, whose lifetimes with respect to hydroxyl removal are approximately 1-3 days under summertime conditions, had decayed so that approximately 60%, 80%, and 90% of the samples, respectively, were below the 3 pptv detection limit. Figure 2 shows that ozone was also very low for these missions, further conf'mning that this was relatively pristine background air exhibiting very little impact Even though wind directions during missions 1, 3, and 4 were similar to those experienced during missions 5, 6, and 7, the winds were lighter, and the area was under less influence from the strong surface high. Figures 2 and 3 show that the concentrations of ozone, C2C14, ethane, ethyne, propane, and benzene were at least a factor of 2 higher than for the clean marine air. Ozone is a good indicator of aged continental pollution but can be perturbed by small additions of highconcentration stratospheric air, particularly at high altitude. Ethyne, benzene and C2C14, are more reliable indicators of relatively recent anthropogenic input. Ethyne and benzene are products of fossil fuel usage and have summertime atmospheric lifetimes of approximately 7 and 5 days, respectively, and C2C14 is a widely used solvent and degreasing agent with a 1-2 month summertime lifetime . These lifetimes are long enough to permit the detection of these species at a considerable distance from their sources, but short enough that the most recent emissions are significant and do not represent relatively small increments superimposed on high background concentrations, as is frequently the case for CFC compounds.
Five-day back-trajectory analysis suggests that the air encountered in missions 1, 3, and 4 had come from a westerly direction, with the North American continent being the probable source region. This conclusion is reinforced by the different characteristics exhibited by the results for missions 1, 3, and 4, compared to the continental air of European origin. For example, Figure 5 shows that the concentration of ethane in the MBL is significantly higher in this air mass than for the other three air masses shown. Trajectory analysis and the relatively low median concentrations of other pollutant gases (e.g., C2C14 and ethyne) compared to European outflow indicate that substantial dilution with clean marine air and photochemical aging must have taken place for at least 5 days. Figure 5 shows that for missions 1, 3, and 4 the median concentration of propane is also elevated relative to ethyne, even though propane is shorter lived than ethyne. In addition, the species ratios of both ethane and propane versus benzene in this air mass are 2-3 times as high as in the two other continental air masses studied. Ethane and propane are emitted as components of natural gas, which is their principal northern hemispheric source [Blake et al., 1994]. Thus, relatively high mixing ratios of ethane and propane are consistent with natural gas emissions from North America making a characteristic impact on the air mass sampled in missions 1, 3, and 4.
Logarithmic ratios of the hydrocarbons in the different air masses also serve to differentiate them. Figures 6a and 6b

from the relative rates of reaction of HO with the various hydrocarbons, and argued that the difference between calculation and observations is the consequence of dilution by air masses of different photochemical ages.
As discussed by Parrish et al., Figure 6a shows that samples with the most recent anthropogenic input have higher nbutane/ethane and propane/ethane ratios, while those measured in the clean marine atmosphere tend to have the lowest ratios. This is expected if the initial ratios are all similar and prolonged exposure preferentially removes the more reactive components. However, it is interesting that the values for the continental air mass of North American origin appear to form a distinct group below the least squares fit line to all the data, with nbutane/ethane ratios as low as the lowest clean marine samples but propane/ethane ratios in the same range as the other two continental air masses. The small range covered by the Continental (North American) data is remarkable, considering that the samples were collected on three different days nearly a week apart. The low n-butane/ethane concentration ratio for these measurements is most likely to have been caused by the significantly larger contribution from natural gas and the longer transit time from North America to the sampling location. However in many cases the European influenced samples were enhanced in n-butane. This is most likely the result of reduced transport time to the experimental region and extensive use thrtughout Europe of liquified petroleum gas which is composed mainly of propane and butanes . A similar effect is seen in Figure 6b for i-butane/ethane versus n-butane/ethane. The least-squares fit through the ibutane/ethane versus n-butane/ethane plot gives a slope of approximately 0.98, which is consistent with previous measurements [Jobson et al., 1994]. Both n-butane and i-butane have similar sources, and the rate of reaction of n-butane with I-• is only 10% faster than for i-butane [Atkinson, 1990]. Therefore, the samples from missions 1, 3 and 4 now provide the lowest ratios on both axes and follow the least-squares fit through the points from the other three air masses. In summary, even though previous comparisons have found remarkably good agreement between hydrocarbon ratios from different air masses, this work emphasizes that the strictly linear relationships and photochemical age progression discussed by, e.g., Parrish et al. [1992], are sensitive to the basic assumption that the air masses have similar initial compositions.
Ethane and propane have slightly higher mean concentrations in the free troposphere than in the M•BL (Table 2). Figure 3 shows that this trend is particularly apparent for missions 5, 6, and 7, the clean marine Lagrangian series of flights. Figure 7 includes the vertical distribution of ethane and ozone for these three flights, and shows that above about 1000 m both of these gases were significantly enhanced. The low concentrations of shorter-lived gases such as ethyne (see Figure 7), n-butane, and n-pentane showed relatively little recent continental influence. Trajectory analysis suggests that at higher altitudes in the free troposphere, missions 5, 6, and 7 may have had some influence from air of North American origin. It has been seen that the flights most likely to characterize emissions from North America (1, 3, and 4) also exhibited particularly high ethane and propane mixing ratios. Thus the relatively high concentrations of ethane and propane observed in the free troposphere for missions 5, 6, and 7 are consistent with some mixing with air of North American origin.  Figure 7 (missions 5, 6, and 7) shows that the MBL ozone is only about half the FT value. This is consistent with a highaltitude stratospheric ozone source and a low-altitude sink. At low altitudes the relatively high moisture associated with clean MBL air is expected to produce net photochemical ozone loss during daylight under low NO x conditions. However, Table 2 shows that the median ozane in the MBL was only slightly lower than for the FT. This reflects the prevalence of photochemical ozone production in the M3L caused by low-altitude European continental outflow into the NE Atlantic.  CFC-12 emissions from Europe during 1987-1990. Figure 4 also shows that this layer contained enhanced methyl bromide concentrations, supporting reports of continental sources for this gas which include agricultural [Yagi et al., 1995] and domestic fumigation and leaded gasoline [Baumann and Heumann, 1987]. No specific trajectory analysis is available, but the meteorological analysis suggests that this tongue of continental air had moved into the area from the Iberian Peninsula [Bretherton and Pincus, 1994]. The same aerosol-rich air mass was sampled the previous day when there was evidence of similar gas enhancement between 1000 and 1500 m. Ethene and propene are significanfiy elevated in the MBL (see Table 2) but do not show significant enhancements in either continental air or in urban plumes (e.g., see ethene in Figure 4). These gases are known to have large urban sources [Singh and Zimmerman, 1992] but are too short lived (atmospheric lifetimes less than 1 day)to have been transported from the continents into the AS2EX/MAGE area. They have been reported to have significant oceanic sources, which likely contributed to the MBL enhancement [Donahue and Prinn, 1990;Rudolph and Johnen, 1990]. However, the prof'fie in Figure 4 shows an example of one of the ethene-enhanced samples or "spikes" which were regularly observed. These spikes were not enhanced in any of the other measured gases except for propene, or occasionally ethyne, which is a good indicator of combustion. These spikes of relatively freshly emitted material may therefore have originated from commercial aircraft and/or ships in the area, or those talcing part in the experiment. Such problems were particularly apparent in the clean marine f'u-st Lagrangian experiment, during missions 5 and 6 which saw intensive use of ship and aircraft sampling platforms, with large spikes superimposed on a very low background.

Most
Mixing ratios of CH3I in the MBL are approximately double those in the FT (see Table 2). For comparison, Table 2 shows CH3Br to be enhanced by approximately 5% in the IVI•L. Figure  5 shows that both halocarbon gases are significantly elevated in air with European continental back-trajectories, with methyl iodide being enhanced by about 30% above clean marine conditions, and methyl bromide by about 20%. However, methyl iodide exhibits less correlation with urban gases in the individual continental European plumes than methyl bromide (see Figure 4). This evidence suggests that a significant component of the methyl bromide may originate from urban areas. It is also consistent with the area to the east of the Azores being particularly rich in nonurban methyl iodide sources. The vertical profiles shown in Figure 7 for the clean marine missions (5, 6, and ?) reveal particularly strong MIlL enhancements for methyl iodide, further implying an oceanic source. Eighty surface samples were collected at different locations and elevations on the Elands of Santa Maria and S•o Miguel. All island samples exhibited elevated levels of CH3I with some being an order of magnitude higher than the 0.6 pptv mean boundary layer value given in Table 2. In addition, concentrations of many other halogensted gases other than CH3Br were substantially elevated in the island samples. These airborne and island observations are consistent with previous work showing higher concentrations of CH3I in North Ariantic coastal waters [Moore and Tokarczyk 1993] and on the English coast in air from the Atlantic [Oram and Penkett, 1994]. Strong convective events are able to rapidly transport this CH3I to the upper troposphere and even the lower stratosphere, particularly in the tropics [Pickering et al., 1992;Kritz et al., 1993]. It has been suggested [Solomon et al., 1994] that short-lived iodine containing gases could play a role in ozone depletion in the tropical lower stratosphere.

Conclusion
During ASTEX/MAGE, circulation patterns resulted in the transport of continental air from North America and Europe to the NE Atlantic. Each of these air masses was characterized by its d•stinctive halocarbon and hydrocarbon composition. The continental air from both North America and Europe was enriched in hydrocarbon and halocarbon photochemical ozone precursors and ozone, suggesting that photochemical production from emissions originating from both continents contribute to the summertime tropospheric ozone budget in the Atlantic region.
European influence resulted in preferentially enhanced MBL concentrations of many anthropogenic hydrocarbons, halocarbons, and ozone. Longer range transport from North America was associated with a more even distribution of pollutants between the free troposphere and MBL. Industrially emitted gases and solvents were used effectively to verify calculated trajectories. Methyl iodide and methyl bromide concentrations were enhanced in the MBL and in air with European trajectories, with some evidence of the influence of anthropogenic/urban sources for methyl bromide. The distribution of methyl iodide, however, was more consistent with oceanic and, in particular, coastal sources playing a dominant role.