Summertime measurements of selected nonmethane hydrocarbons in the Arctic and Subarctic during the 1988 Arctic Boundary Layer Expedition (ABLE 3A)

,


this issue (a)].
Approximately 1150 tort of air from each sample canister was transferred to a 2-L stainless steel storage vessel, which was fixed permanently to the vacuum line.The hydrocarbons were trapped using a vacuum pump to pull the air sample through a preconcentration loop (8 inch x 1/4-inch O.D. stainless steel) filled with 1-mL-diameter glass beads and irmnersed in liquid nitrogen.After 1000 torr of the sample had passed through the glass beads, the trap was isolated and heated with warm water.Then the helium carrier was redirected to flush the previously trapped contents from the loop to the separation column.The temperature of the column then was a series of later experiments were carried out on the Alaskan samples at UCI with an A120 3 porous layer open tubular (PLOT) column (Chrompak) which provided a clean separation of these two C 5 molecules, as shown in Figure 2. When very low concentrations were observed for C3H 8 and n-C4H10 in the atmospheric samples and isoprene concentrations were elevated to the ppbv range, n-pentane was a very small interference (< 5 pptv).While such n-pentane/isoprene reanalyses were carried out on numerous Alaskan air samples, most of them were no longer available for reanalysis.Therefore, in the samples analyzed only in Alaska on the GSQ column, large isoprene/n-pentane peaks were assigned as isoprene when accompanied by n-butane near the detection limits.With slightly higher n-butane mixing ratios, the npentane mixing ratios by inference represented only a minor fraction of the combined peak and were assigned from the nbutane/n-pentane ratios in isoprene samples collected well outside the planetary boundary layer.
The precision of the air sample measurements were estimated to be as follows: Ethane, ethene, ethyne, propane and propene (3%); n-butane, /-butane, and i-pentane (8%); and n-pentane/isoprene (10%).The qualitative detection limit for the C2-C 4 hydrocarbons was estimated to be about 2 pptv, while the detection limit of the C 5 hydrocarbons was estimated to be about 10 pptv.

Concentration C:Ha C2H6
apparatus allowed approximately 50 air samples to be analyzed per 24-hour period.
The gas chromatograph was interfaced to a Spectra Physics 4270 computing integrator and a PC computer for data storage and reduction.While the computer-controlled data storage was adequate for preliminary data output, the variations in baseline behavior and retention times required subsequent visual examination of each chromatogram.
A clean, partially dried (3 ppmv H20 ), whole air sample contained in an Aculife-treated Luxfer cylinder pressurized to about 2000 psi was used throughout this project as a working standard.It was assayed after every eight air samples in the same manner and procedure used to analyze the samples.Repetitive comparisons with several gas mixtures employed as reference standards demonstrated that the selected gases in this working standard exhibited no statistically significant changes in their mixing ratios over the course of the ABLE 3A project.Reference standards with a stated accuracy of 2% (95% confidence limit) were obtained from Scott Specialty Gases (Plumsteadville, Pennsylvania).The mixing ratio of propane in the Scott standard was in turn confirmed by comparison to a methane-propane mixture provided by the National Bureau of Standards (SRM 1660A) with a stated accuracy of 1% (95% confidence limit).We have assumed the FID response for isoprene (C5H8) and n-pentane (C5H12) to be the same.The accuracy of the calibration procedure is estimated to be 5%.
Although the GSQ column utilized during the Alaskan phase of measurement did not resolve n-pentane and isoprene,

RESULTS: OVERVIEW OF FLIGHTS
More than 1000 air samples were collected during the 33 missions flown during this GTE project.The typical collection patterns produced series of air samples representing flights at various constant altitudes and of vertical profiles from ascending or descending tight spirals over fixed points.Figures 3 and 4 show flight paths and campaign areas and Table 1 summarizes information for each vertical profile made.Commencing in early July, the first four missions constituted the transit from Wallops Island, Virginia (37.9øN, 75.5øW), to northern Alaska via Thule, Greenland (76.6øN, 68.8øW).Samples were collected in the middle free troposphere between 4500 and 6000 m to furnish the latitudinal and longitudinal profiles shown in Figures 5 and 6 and data given in Table 2.In these and all subsequent figures, the mixing ratios for isoprene are only reported if one or more of the samples showed concentrations sufficiently elevated over background to insure that the chromatographic peak was not plausibly attributable to n-pentane.No such samples were encountered prior to flights over the Alaskan boreal forest.
One vertical profile was made near Thule (Figure 7).On The Arctic and sub-Arctic vertical profiles revealed background planetary boundary layer mixing ratios for ethane, ethyne, propane and n-butane of 819+_42, 49+13, 49+12, and 8+_4 pptv, respectively, while the mixing ratios at altitudes of greater than 4 km were 891+95, 74+23, 66+30, and 9+8 pptv, respectively.An increase of mixing ratio with altitude was observed during all but two missions for C2H 6, C2H 2 and C3 HS.The mixing ratios of n-C4 H !0 were generally so close to the detection limit that trends with altitude are of marginal statistical significance.

Hydrocarbon Source Signatures
More than 90% of the air samples contained mixing ratios of the four hydrocarbons, ethane, ethyne, propane, and n-butane, which fell within a rather well-defined pattern, e.g., ethane between 600 and 1200 pptv.Much of the variation within this group can be assigned to temporal and altitudinal variation, i.e., lower mixing ratios at lower altitudes and later in the summer.Within this generalized background, it is difficult to isolate any specific input sources.However, among the remaining 10% of the air samples, some easily noticeable mixing ratio variations are observable which identify specific NMHC sources.The most obvious source is the boreal forest, with isoprene as its unique marker.The atmospheric lifetime of isoprene against HO reaction is only an hour or two, and the marker is therefore not found very far from its local source.The experimental data show that isoprene emission is observed only on flights over or adjacent to the boreal forest, demonstrating by its absence that neither the polar ocean nor the tundra are isoprene sources.First, natural gas samples were collected from the Alaskan North Slope area.The NMHC composition of natural gas is known to be variable, depending on whether it is derived from the biogenic decomposition of organic matter or is associated with crude oil or coal deposits.These Alaskan samples had a particularly low concentration of NMHC compounds (99.5% CH 4 by volume of total hydrocarbons) and contained 0.4% ethane, 0.05% propane, 0.02% n-butane, and no measurable ethyne (<0.001%).
Anchorage, the largest urban area in Alaska, was chosen to represent Arctic and sub-Arctic cities.Four samples collected in July upwind of Anchorage yielded average background mixing ratios of 1770 ppbv for methane and 750, 50, 50, and 30 pptv of ethane, ethyne, propane, and n-butane, respectively.These values were subtracted from the average mixing ratios of 12 samples collected at locations throughout the city to yield average excess mixing ratios of approximately 150,000, 750, 2250, 150, and 450 pptv for methane, ethane, ethyne, propane and n-butane, respectively.From these values, the urban NMHC ratios relative to ethane reported in Table 4 were calculated.The several-fold higher excess contribution from ethyne versus ethane is known to be particularly characteristic of fossil fuel combustion.
Furthermore, sets of air samples were collected on two separate occasions upwind and downwind of the oil drilling activities located near Prudhoe Bay, Alaska.The upwind samples yielded average background concentrations of 1780 ppbv for methane, and 802, 60, 71, and 49 pptv for ethane, ethyne, propane, and n-butane, respectively.The downwind samples yielded average mixing ratios of 1785 ppbv for methane and 1408, 181,492, and 295 pptv for ethane, ethyne, propane, and n-butane, respectively, from which the excess ratios given in Table 4 were calculated.The precision of the C H 4 measurements is estimated as +4 ppbv, making the downwind concentrations marginally higher than the upwind values.The other gases were clearly augmented on the downwind side.The Prudhoe Bay samples had the lowest ratios relative to ethane for methane and ethyne (8+4 and 0.2) and the highest propane ratio (0.7) of any sampled source.13).Similarly, excesses of these hydrocarbons are found at 4000 m in both profiles of mission 20 (Figure 16) and the first profile of mission 21 (Figure 17).The assignment of "background" hydrocarbon concentrations is an important aspect in the evaluation of the "excess" hydrocarbon concentration ratios presented individually in Table 5 and in summary in Table 4.The error estimates shown in Table 5 have been evaluated from the standard deviations of a linear fit to the background air samples.For example, four samples above and four below the fire enhanced layer samples collected during the descending spiral of mission 20 were employed to evaluate the background concentrations and errors.Of the measured NMHC, ethane and ethyne exhibited the greatest enhancements from biomass burning and were remarkably uniform between different plumes.These two NMHC also have the longest atmospheric lifetimes and would be expected to be the least diminished during the time elapsed between emission from the fire and collection of the samples.
The propane and n-butane excesses were relatively small so that n-butane could not provide reliable enhancement ratios, only being reported as less than 0.01.This is consistent with the fact that propane and n-butane are shorter lived (and, by the same reasoning, would be expected to have lost a substantial fraction of their initial concentrations during the transit time between emission and sampling, estimated from trajectory analysis to be 25-50 hours).The correspondingly lower mixing ratios of these gases also make it more difficult to assign baseline values.These factors were accordingly taken into account in the error calculation.Thus, it is suggested that the range of values given in Tables 4 and 5 4).Both papers quote somewhat higher propane and lower methane ratios to ethane.Two other prominent regions of enhanced hydrocarbon concentrations can be identified in the vertical profiles for mission 23 near 4000 m (Figure 19, descending data only) and for mission 29 near 3000 m (Figure 23).The relative excesses of the hydrocarbons from these two profiles are calculated in Table 6.The data from profile 23D were collected from an air mass over Bristol Bay, while the vertical spiral on mission 29 was over the northern e:lgc of Greenland and completely dissociated from possible Alaskan forest fire contributions.NMHC in the region northeast of Prudhoe Bay (Figure 22).The excess hydrocarbon ratios relative to ethane for each of the two samples collected in this region (153øW -145øW) (see Table 6) show that they represent two distinct air masses.
Comparison of these hydrocarbon ratios with the source ratios of Table 4 indicates  The average mixing ratios -I-lo for the latitudinal range covered by the second latitudinal survey during the return to Virginia were 980+120, 105+30, 150_+60, and 6+_2 pptv for ethane, ethyne, propane, and n-butane, respectively (Figure 21 and Table 2).The northernmost air mass encountered (between 78 ø and 83øN) was suggested to be of a single Arctic origin, with the probable exception of the air mass containing enhanced hydrocarbon 6oncentrations at the 3000 m level in the vertical profile (Figure 23 and Table 6).Consequently, only minor variations in the concentrations of the gases were observed (Figure 21).For the remainder of the flights, the air masses were of stratospheric and/or continental origin and showed much greater variability.The observed variation in the mixing ratios of the selected gases displayed in Figure 21 supports the trajectory analyses.
In conclusion, neither of the latitudinal surveys, which were performed 6 weeks apart in this summertime experiment, revealed any statistically significant dependence of the measured hydrocarbons within this range of latitude.This observation is consistent with the latitudinal profiles that have been observed in surface samples collected during the same season at similar latitudes by Blake andRowland [1986 andunpublished data, 1990].Table 2 shows that the mixing ratios are within and mostly at the low end of the ranges observed by other studies made in this latitude band in summer

Vertical Profiles
Almost every mission flown during the ABLE 3A project contained at least one vertical profile.In all, 37 spiral vertical profiles were made between 150 and 6200 m (Table 1).Various profile locations were chosen in order to examine unpolluted regions of the lower troposphere and to investigate regional phenomena such as emissions from biomass burning, longrange transport of pollutants, and emissions from tundra and   ,,,i,,,,i,,,,i,1,,i,,,,i,,,,i,,,,i,, The frequent observation of very high correlation coefficients among hydrocarbons in a vertical profile, with higher concentrations at higher altitudes, indicates the operation of a rather consistent mechanism for simultaneous alteration of all of the mixing 'ratios.For all four of these hydrocarbons, the primary atmospheric sink is reaction with HO radical, as in (1) and ( 2 In the actual atmosphere, however, these completely isolated, constant temperature conditions do not hold.In addition there will be some contribution from dilution. Nevertheless, the measured slopes of the logarithmic plots are not too different from the calculated rate constant ratios in Table 7 for C2H 2 and C3H8, and all four logarithmic hydrocarbon loss rates fall generally in the same relative order characteristic of the rate constants at the lower altitudes.The very high correlation coefficients strongly suggest that all of the hydrocarbons (other than isoprene) in these air masses at different altitudes in a vertical profile tend to have a generic common origin subjected later to diminution in mixing ratios by HO attack.This common origin is primarily assumed to be release of the hydrocarbons in the inhabited areas of the temperate latitudes.At these lower•atitudes, the air masses generally show decreasing mixing ratios of hydrocarbons with increasing altitude, so that the subsequent reversal to the opposite tendency requires even more hydroxyl processing than if the starting profiles had exhibited no change with altitude.* For all alkynes.

The formation of tropospheric HO is initiated by ozone absorbing ultraviolet radiation
travel more rapidly from the source region than those from lower altitudes, then less exposure time to HO attack could have accumulated upon reaching a particular Arctic latitude and longitude, and the initial hydrocarbon ratios would be more nearly intact.Third, another variant of this possible differential exposure to hydroxyl is the suggestion of subsidence from higher to lower altitudes.This scenario invariably requires longer exposure to reach lower altitudes and, therefore, greater hydrocarbon losses.This third explanation would begin with rapid upwelling into the upper free troposphere from the source region by some fast mechanism such as cloud pumping [Gidel, 1983].As it moved aloft, a general subsidence of some portion of the air mass and subsequent mixing with cleaner low-altitude air could occur.Supporting this possibility is the argument that highaltitude, long-range transport of PAN, followed by subsidence and subsequent thermal decomposition, transports a significant fraction of NO x from urban sources areas to Arctic and sub-Arctic regions [Singh et al., this     was highest, decreased with increasing altitude for both profiles (Figure 14), from about 800•_40 pptv to 710 pptv on the first and 740 pptv to 680 pptv plus an alkane plume at 1500 m on the second.The mixing ratios for ethyne also decreased toward the higher altitudes, but n-butane at generally quite low mixing ratios showed the opposite behavior, with higher values at the higher altitudes.The propane data showed more variation, with several higher values well correlated with apparent excess ethane.These incremental excess propane concentrations were more significant on a percentage basis than similar additions for ethane because the background concentration of the latter was about twelve-fold higher than for propane.These incremental amounts of propane make its underlying correlation with altitude uncertain as to sign, but small in magnitude.

Boundary Layer Measurements
The very high NMHC levels observed at lower altitude (in the PBL) during mission 33 are of interest.With the exception of isoprene and ethane, Table 8 shows that the mixing ratios of each NMHC were less at 135 m than at 330 m (see also Figure 27).Hydrocarbon oxidation in the presence of sufficient NO x leads to the formation of tropospheric ozone where either the hydrocarbon or the NO x concentration will be the factor limiting the total ozone production.However, ozone formation in a specific region is more strongly dependent on the reactivity of the hydrocarbons that are present, rather than the total carbon concentration.This is especially true of isoprene, which is highly reactive toward HO radicals, and thus is a species of particular relevance to localized ozone production.The reaction rate constant for HO with isoprene is about a factor of 100 faster than for propane, so that lifetimes of only an hour or two are typical for the former.These missions were interesting in terms of the welldefined biomass burning plumes, and also because the boreal forest area overflown on the descending spirals was populated by known isoprene emitters, such as white and black spruce, paper birch, balsam poplar, and aspen [Rasmussen, 1970].

Fig. 1 .
Fig. 1.Hydrocarbon chromatogram with the GSQ column for a sample collected in the planetary boundary layer over a boreal forest located in central Alaska.

Fig. 2 .Fig. 3 .
Fig. 2. Hydrocarbon chromatogram with the Al203 PLOT column for a sample collected in the planetary boundary layer over a boreal forest located in central Alaska.

Fig. 4 .
Figure 8 (and see Table 3).The remainder of the data are shown individually in Figures 9 to 11. Next, the focus of the research flights shifted to the sub-Arctic regions around Bethel, Alaska (60.8øN, 161.8øW)where many similar vertical profiles were made over the boreal forest, tundra, and the northern Pacific Ocean.A composite profile representing sub-Arctic background conditions (six profiles of missions 13, 15, 17, and 25) is shown in Figure 12.The remaining profiles made near Bethel are shown in Figures 13 to 19, and Figure 20 shows a composite of all the Arctic and sub-Arctic background vertical profile data (Table 3).The project was concluded with the return transit to Wallops Island made in mid-August (missions 28-33) when further longitudinal, latitudinal, and vertical profile samples were collected, the results of which are shown in Figures 21 to 27.Further details of the individual flights and their categorization can be found in subsequent sections of this paper and in the overview paper of Harriss et al. [this issue (b)].All references to meteorological data and air mass trajectories are derived from Shipham et al. [this issue].All of Fig. 5. Latitudinal profile of ethane, ethyne, propane, n-butane (in Fig. 6.Longitudinal profile of ethane, ethyne, propane, n-butane (in pptv), and longitude (in degrees west) collected during transit missions 1-pptv), and latitude (in degrees north) collected during transit mission 4, 3, flown between Wallops Island, Virginia, and Thule, Greenland, on July flown between Thule, Greenland, and northern Alaska on July 9, 1988.7-8, 1988.
type of vegetation burned.Greenberg et al.  [1984]  measured relatively high ethyne emissions from Amazonian fires, but the emission of ethyne reported by Hegg eta/.[1990]  for the burning of forests and brush in the United States and Canada were more consistent with our Alaskan findings (Table
The reactions with these hydrocarbons of other atmospheric oxidants are much too slow to play significant roles in their atmospheric removal[Atkinson, 1990].With an isolated air mass, the concentrations of all four hydrocarbons can be expected to diminish logarithmically versus integrated exposure to the atmospheric level of HO.
energetic enough to produce O(•D) atoms by (3) which then react with water vapor by (4) to form two hydroxyl radicals.The yield of HO from O(•D) is issue; Jacob et al., this issue; Sandholm eta/., this issue].While attractive as an explanation for NOxtransport , this meteorology seems quantitatively incapable of producing the large concentration changes observed in most low-altitude air masses.Our data do not provide any definitive resolution among these meteorological mechanisms, nor is it likely that a single meteorology suffices.Whatever the meteorology, however, the dominant factor in explaining the strong correlations is chemical removal of the hydrocarbons by reaction with hydroxyl radicals.The ABLE 3A missions can be compared with a number of vertical profiles from the STRATOZ H and STRATOZ HI projects [Ehhalt et al., 1985; Rudolph, 1988].Those experiments were flown in June of 1980 and 1984 and covered a latitudinal range from 70øN to 60øS at altitudes up to 12 km above the Atlantic Ocean.The longitudes involved in STRATOZ were those of Greenland and the east coast of NorthAmerica.Their measured northern hemisphere concentrations were somewhat higher than presented here (Table2).The profiles measured at mid-northern latitudes and throughout the southern hemisphere exhibited a general decrease in the mixing ratios of the selected NMHC with increasing altitude but, interestingly, both STRATOZ H and IH showed the opposite trend at high northern latitudes.This increase of NMHC concentrations with altitude was interpreted in terms of fast mechanisms of upward transport such as the cloud pumping suggested byGidel [1983].The same phenomenon also has been observed in the Norwegian Arctic by Tille et al.[1985], and hydrocarbon "bulges" above the PBL have been reported in the region of Alaska by Rasmussen eta/.decrease in the levels of ethane, ethyne, and propane with altitude.Several similar vertical profiles were made in and around the Bethel region, but mission 17 was Fig. 29.Ln-ln plot for ethyne, pr.op.ane, and n-butane versus ethane for the vertical profile made during nuss•on 30.

Fig. 28 .
Fig. 28.Ethyne, propane, and n-butane (in pptv) versus ethane for the Fig. 30.Correlation of mixing ratios of ethane and ethyne for nonplume vertical profile made during mission 30.air samples of missions 20 and 21.
from these observations that the air sampled during both spirals of mission 17 was quite different from all other Arctic and sub-Arctic air masses studied during the ABLE 3A project.Significantly, the decrease in the concentrations of NMHC with increasing altitude was accompanied by a unique trajectory indicating that air masses of southeastern origin (i.e., from the North American continent) are quite distinct in character from those that originate over either the Asian continent or the north-central Pacific.Among other possible factors is the likelihood of shorter exposure times to HO radicals at the lowest altitudes.The only other profile that did not conform to the general pattern of increasing mixing ratio with altitude was made during the final mission (33), on August 17, 1988.The Electra passed over weftands in the Wallops Island area and provided both the southernmost vertical sounding and the air mass with the highest NMHC concentrations• sampled during the entire project.Altitude versus mixing ratio plots for ethane, mass was augmented by natural gas emissions.The ethyne/n-butane correlation (r • = 0.92) and elevated levels of ethane and propane imply strong combination influence.The air mass is further enhanced by local isoprene emissions.The negative correlations of the NMHC species with altitude displayed in Figures 14 and 27 are thought not to be solely the result of enhanced hydrocarbon levels at the surface, because these are coupled to unusually low concentrations in the air mass at the top of the profile.The air mass encountered above 3 km on mission 33 was found to be particularly low in NMHC and was significantly different from that encountered at similar altitudes two days earlier during mission 30 (Figure 24).For example, the mixing ratios of ethane and ethyne of 884 pptv and 38 pptv, respectively measured at 5300 m during mission 33 were substantially lower than the 1100 pptv of ethane and 125 pptv of ethyne observed two days earlier during mission 30 at the same altitude.In addition, the methane levels observed during the high-altitude portions of missions 30 and 33 were 1660 and 1590 ppbv, respectively [Harriss et al., this issue (a)], and a northern hemispheric methane mixing ratio for methane of only 1590 ppbv is low enough to imply that the air mass was of either stratospheric or tropical origin.The observations of low NMHC mixing ratios alone are consistent with both these possibilities; however, the relatively low in situ ozone data measured in this region [Gregory et al., this issue] (Figure 27) considerably favor the case that the air sampled at high altitudes was of tropical origin.This conclusion is also supported by the available trajectory data [Shipham et al., this issue].
sub-Arctic areas explored during ABLE 3A were significantly influenced by long-range transport of NMHC from distant source regions.Trajectory analysis suggests that the only high-latitude air masses that did not conform to the normally observed profiles were of a southern origin.Plume enhancement of some or all of the measured NMHC were observed on more than half of the 33 missions flown during this GTE project.The usual summer vertical profile of reactive hydrocarbons in these high latitudes has elevated concentrations at high altitudes, with mixing ratio variations largely controlled by hydroxyl radical reactions.Because of the very dry conditions prevailing during the project, wildfires were widespread and were established as a significant source of numerous NMHC.Biomass burning emission factors relative to ethane were established for ethyne (0.38+0.04) and propane (0.08+0.03).Activities associated with oil drilling are a probable source of the enhanced levels of alkanes that were obaerved as much as 300 km northeast of Prudhoe Bay, Alaska.Isoprene emissions from boreal forest Fig. 31.Modeled vertical profile of ozone under conditions regions of the Yukon Valley dominate local photochemical HO corresponding to mission 33 in the case of high and low isoprene reactions.Biogenic production of isoprene in eastern emissions (I. S. A. Isaksen, personal communication, 1991).Virginia, coupled with long-range transport of altitude or in the transition from the boreal forest region to the wetland area.
Figures 16 and 17 show that the mixed samples collected in this forest region contained isoprene mixing ratios between 400 and 550 pptv at altitudes up to 300 m on mission 20 and up to about 1000 m on mission 21.By contrast, the samples collected at similar altitudes in the wetlands area near Bethel exhibited isoprene mixing ratios in the 50-120 pptv range.This observation is consistent with the fact that the moist tundra regions contain a much lower population of isoprene emitters.The first spiral, made at 1000 local time over the boreal forest, yielded an isoprene profile which shows detectable concentrations of isoprene up to approximately 2000 m.Six hours later, during mission 21, isoprene was observed up to 2500 m.The available meteorological data for missions 20 and 21 indicate that the top of the mixed layer was between 1800 and 2000 m for mission 20 and between 1800 and 2400 m during mission 21.These observations are consistent with the fact that isoprene, with an atmospheric lifetime of an hour or two during daylight hours, is not expected to persist long enough to escape the PBL and be observed in the free troposphere.This is a good example of how vertical profiles of short-lived gases released at the surface can be used to corroborate meteorological analyses.In addition, and notably unlike mission 33, Figures 16 and 17 show no ozone enhancement caused by the oxidation of isoprene.In fact, Gregory et al. [this issue] concluded that the fast reaction of isoprene with ozone was an efficient sink for ozone in the region of boreal forests.The major difference in the near-surface chemistry of missions 20 and 21 versus mission 33 lies not in the isoprene concentrations which are quite elevated in the PBL in all flights, but in the greatly enhanced NO x levels in mission 33.The rate-determining step for formation of tropospheric ozone from isoprene emissions anthropogenically produced nitrogen oxides from urban areas, is responsible for significant production of regional tropospheric ozone.Acknowledgments.The authors greatly appreciate the following research group members for their assistance in the fabrication of the experimental apparatus, comments regarding the manuscript, and support during the field mission: K.C. Clemitshaw, T. Gilpin, N.R.P. Harris, R. lyer, T. Merrill, P. Russell, M. Schaeffer, J. Wang, and N. Wang.I. S. A. Isaksen also commented extensively on both the observations and interpretation of the data, and D. H. Ehhalt made some useful suggestions.J.P. Iwanicki, K. Kaufman, and M. Mastropaolo participated in the analytical processing at the University of Alaska-Anchorage throughout the mission.We are indebted to Ralph Kolbush, director of the UCI machine shop, for excellent support before, during, and after the missions.The opportunistic filling of air canisters when an interesting atmospheric air parcel was encountered by the Electra was dependent upon the timely relay of information from the operators of real-time, on-board instrumentation.We especially thank the personnel of the University of Alaska-Anchorage for the use of laboratory space and for frequent assistance with items in logistical support of the analytical laboratory.Ed Clinton of Alaska Valve and Fitting Company was particularly generous and a valuable source of spare parts for the chromatographic apparatus.This work was supported by funds from the National Aeronautics and Space Administration grant NAG-1-783.