Influence of southern hemispheric biomass burning on midtropospheric distributions of nonmethane hydrocarbons and selected halocarbons over the remote South Pacific

,


Experiment
PEM-Tropics A employed two aircraft, the NASA Ames DC-8 and the NASA Wallops P-3B, and took place between August 24 and October 6, 1996.The payloads of both aircraft included instruments capable of quantifying an extensive suite of atmospheric species, including NMHCs, CH3C1, C2C14, alkyl nitrates (RONO2), CO, 03, NOx, peroxyacetyl nitrate (PAN), nitric acid (HNO3), and aerosols [Hoell et al., 1999].The two aircraft sampled an area of the Pacific stretching from northeast of Hawaii to the tip of Antarctica (45.2øN-72.5øS),and northwest of Fiji to the west coast of South America (152.7øE-78.0øW)(Plate 1).The scientific payloads of the two aircraft differed in some respects, reflecting the operating characteristics of the two platforms.The high-altitude ceiling (typical altitude range 0.3-11.9km during PEM-Tropics A) and long range of the DC-8 favored the extensive geographical survey of the South Pacific troposphere, a region for which few trace gas data were available previously.By contrast, the low-altitude capabilities of the P-3B (typical altitude range 0.06-8.5 km during PEM-Tropics A) allowed detailed characterization of Marine Boundary Layer (MBL) processes.
As part of the International Global Atmospheric Chemistry (IGAC) project First Aerosol Characterization Experiment (ACE 1), the National Center for Atmospheric Research (NCAR) C-130 aircraft flew an extended latitudinal survey between 76øN and 60øS through the center of the Pacific basin during early November 1995 (Plate 1).A maximum flight altitude of 6.2 km was attained by the NCAR C-130 in the Southern Hemisphere during this series of southbound transit flights.
The University of California, Irvine (UCI), group collected whole air samples aboard the three aircraft during PEM-Tropics A and ACE 1 using procedures described by Blake et al. [1992Blake et al. [ , 1994]].
The gas chromatographic analytical apparatus employed after the samples were returned to UCI is described by Blake et al. [1996a], Sive [1998]

Source Signatures
Tropospheric 03 has two predominant sources, in situ photochemical production from its precursors (CO, NMHCs, and NOx) and downward mixing from the stratosphere.When combined with trajectory analysis, the contributions of different precursor sources can be assessed using a wide range of tracers for different source types.Therefore the use of chemical tracers plays a crucial role in establishing the most likely origin of O3-enriched layers.
Incomplete combustion is the principal global source of CO, ethane (C2H6), and ethyne (C2H2).Combustion occurs not only in natural wildfires and in association with land clearing and agricultural land use but also in urban areas through fossil fuel consumption [Andres et al., 1996].However, some gases such as C2C14 are specific to industrial/urban activities and serve as markers for such urban

Spatial Distributions
Plates 2-4 are color patch plots showing the latitudinal and longitudinal distribution of the mixing ratios of CO, 0 3 , C2H 6, C2H 2, CH3C1, and C2C14 for PEM-Tropics A. The data were averaged over a latitude-longitude grid size of 5 ø by 5 ø and are divided into three altitude ranges.The vertical bins approximately correspond to the Marine Boundary Layer (MBL)/lower troposphere (0-2 km), the midtroposphere (2-8 km), and the upper troposphere (8-12 km).Plates 5-7 show the corresponding 5 ø by 5 ø patch plots for ACE 1.The ACE-1 data are described by Blake et al. [1999], and are displayed here in the same form as PEM-Tropics A in order to facilitate comparison between the two data sets.However, in the southwestern region (west of about 155øW and between 10 ø and 55øS) they typically were more coherent and contained higher mixing ratios of CO, 03, C2H6, C2H2, and CH3C1.
Figure 1 shows a vertical profile through one layer observed between about 3 and 7 km altitude.The profile was sampled by the DC-8 near Tahiti on September 5, 1996.The layer exhibited elevated mixing ratios of CO, 03, C2H 6, C2H 2, and CH3C1, but C2C14 remained at background levels.Enhanced mixing ratios of photochemical species other than 03, including PAN, alkyl nitrates, nitric acid, formic acid and acetic acid, were also observed in such layers [Schultz et al., 1999;Talbot et al., 1999].

Easterly Outflow From South America
Elevated mixing ratios of gases characteristic of biomass burning emissions, including CO, CH3C1, C2H 6, C2H 2, and C3H8, were observed in all altitude ranges during three local P-3B flights conducted from Guayaquil, Ecuador, September 18 -23, 1996 (Plates 2-4).The mixing ratios of these gases were most elevated over background values in the altitude range 1.5 -4 km (Figure 5).The high values for C2H2/CO (Figure 5) indicate that many of the sampled air masses were relatively fresh, compared to the plumes arriving from the west discussed Regressions for C2H 6, C2H 2, and CH3C1 versus CO in the altitude range 1.5-4 km for the three Guayaquil local flights were highly correlated (Figure 6).The corresponding enhancement ratios for C2H 6, C2H 2, and CH3C1 versus CO (7.4,2.9, and 0.73 pptv/ppbv, respectively) were remarkably similar to those observed in biomass burning plumes during TRACE A [Blake et al., 1996b;Mauzerall et al., 1998].The good correlations for C2H 6, C2H 2, and CH3C1 versus CO, and generally poor correlations for C2C14 versus CO, suggest that the air masses sampled just off the coast of South America were most strongly influenced by biomass burning emissions, rather than by urban emissions.
Numerous backward trajectories for different locations along the Guayaquil local flight tracks traveled over South America [Fuelberg et al., 1999], including regions of Brazil where biomass burning usually is widespread during September [Fishman et al., 1996;Olson et al., 1999].Figure 7  The 03 versus CO regression slopes of 1.5 and 1.9 ppbv/ppbv for PEM-Tropics A and ACE 1, respectively, are in reasonable agreement (Figure 9).However, they are greater than the average O3/CO enhancement ratio of 0.74 ppbv/ppbv calculated by Mauzerall et al. [1998] for approximately one week old biomass burning plumes observed in the South Atlantic region during TRACE A. We recall that many of the South Pacific plumes were aged more than one week so may have been influenced by the photochemical loss of CO, which causes the O3/CO ratio to increase with age [Mauzerall et al., 1998].Net photochemical production of 03 was calculated to continue at altitudes above about 6 km as the plumes were advected over the South Pacific during PEM-Tropics A [Schultz et al., 1999].This source may also have contributed to the high Pacific O3/CO ratios.By contrast, C3H8 in perturbed air masses showed relatively small and insignificant (+l t J) increases above background in the 2-8 km altitude range (Figure 10).The average lifetime of C3H 8 in the tropical midtroposphere is approximately 10 days.Dilution as well as chemical loss, needs to be taken into account when considering aged plumes.However, the lack of significant midtropospheric (2-8 km) C3H8 perturbation supports plume transport times of the order of 2 weeks from biomass burning source regions.The urban tracer C2C14 exhibited no significant differences for CO <55 ppbv and CO >70 ppbv (Figure 10).10).This suggests that on average, the MBL was not significantly affected by net local oceanic emissions of these gases.
Average 0 3 levels were significantly elevated (_+lc•) for CO >70 ppbv compared to CO <55 ppbv in the 2-8 km range (Figure 10).The average 03 mixing ratios for the CO >70 ppbv mixing ratio bin for 2-8 km was 67 ppbv, compared to the background average of 33 ppbv.We recall that the observed midtropospheric CO mixing ratio enhancements over the South Pacific were caused by biomass burning.Therefore biomass burning sources caused 03 levels to be approximately double those for relatively clean background conditions during PEM-Tropics A. In fact, the CO <55 ppbv vertical profile for 03 exhibits a "bulge" in the midtroposphere somewhat similar to that observed at high CO values (Figure 10).This suggests that even "background" 03 may have been influenced by the widespread pollution encountered throughout the midtroposphere of the South Pacific region.
Results from a similar analysis employing the ACE-1 data gave a comparable 50-100% increase in 03 mixing ratios for the biomass-burning-impacted plumes compared to background conditions for the 2-6 km altitude range.
However, fewer data were available for these calculations.

Conclusion
, and B.C. Sive et al. (manuscript in preparation, 1999 (hereinafter referred to as S99)).Both operations are summarized below.Aboard the aircraft, a metal bellows pump was used to fill evacuated sample canisters to a pressure of 40 psi.At the UCI laboratory, a 1520 cm 3 (STP) aliquot from each canister was preconcentrated cryogenically in a sample loop.Hydrogen carrier gas flushed the contents of the loop to a splitter, which partitioned and directed the sample flow to five gas chromatographic columns in a reproducible manner.60 m, 0.25 mm ID DB-5MS column, film thickness 0.5 !,tm (J&W Scientific) were used for the analysis of selected C1-C2 halocarbons and methyl nitrate.An ECD and an RTX-1701 column were employed for the determination of C •-C2 halocarbons and C•-C4 alkyl nitrates.The unique separation characteristics of each column were optimal for a particular subset of NMHC or halocarbon gases andby J. Calvert and E. Apel at NCAR [Apel et al., 1994].Results from NOMHICE demonstrate that our analytical procedures consistently yield accurate identifications of a wide range of unknown hydrocarbons and produce excellent quantitative results [Apel et al., 1994; Sire, 1998; S99].Fast-response in situ measurements of CO were made spectroscopically using a tunable diode laser instrument during PEM-Tropics A [Sachse et al., 1991] and were made using a gas filter correlation analyzer during ACE 1 [Kok et al.. 1998].In situ 03 measurements for both PEM-Tropics A and ACE 1 employed NO+O 3 chemiluminescent detectors (e.g., G. Albercook et al., unpublished data).The sampling frequencies of the in situ CO and 03 instruments were much higher than our whole air sampling times during both PEM-Tropics A and ACE 1.We have employed a merged data file generated at Harvard University and containing CO and 03 mixing ratios averaged over the whole air sampling times for comparison of the different PEM-Tropics A measurement databases.The PEM-Tropics A measurements and merged data files are archived at NASA Langley Research Center and can be accessed via the GTE web site at http://www-gte.larc.nasa.gov/.For ACE 1, a merged data set was prepared at UCI.The ACE-1 data archive is maintained by the University Corporation for pptv), and CH3C1 (560-600 pptv) were observed in the midtroposphere (2-8 km) over the southwestern Pacific during both PEM-Tropics A and ACE 1.For PEM-Tropics A, this enhanced region extended from the southern tip of New Zealand to about 10øS and east to about 155øW.Ozone mixing ratios of as much as 80 parts per billion by volume (the southwestern Pacific was experiencing heavy burdens of biomass burning emissions (CO, C2H 6, and C2H 2, and CH3C1) and secondary pollutants (O3), but no significant urban influence (low C2C14).The southwestern Pacific enhancements observed during PEM-Tropics A originated from the west and were in the form of distinct layers of pollution [Fuelberg et al., 1999; Fenn et al., this issue].Such layers were encountered during all but one flight over the remote South Pacific (west of 100øW).
Figure 10 shows the vertical distribution of trace gases for background and relatively perturbed conditions over the South Pacific during PEM-Tropics A. For this figure, all Southern Hemisphere data collected in the southwestern and central South Pacific (south and east of Fiji to 100øW) have been segregated into bins defined by CO mixing ratios <55 ppbv and CO >70 ppbv.The average and median CO values for the entire data set were 70 and 71 ppbv, respectively.A histogram distribution showed the CO mixing ratios to be skewed to give a peak mixing ratio of 57 ppbv and cluster of samples with CO greater than about 65 ppbv, representing the numerous biomass burning layers described earlier.Samples with CO mixing ratios >70 ppbv were chosen to represent air

Figure 10 .
Figure 8.Ten day backward trajectories arriving at 250 hPa along the northerly part of the PEM-Tropics A DC-8 flight 8 track at 1200 UTC September 10, 1996.Small arrows denote locations at daily intervals; large arrows denote locations at 5 day intervals.