Tropospheric hydroxyl and atomic chlorine concentrations, and mixing timescales determined from hydrocarbon and halocarbon measurements made over the Southern Ocean

. During the First Aerosol Characterization Experiment (ACE 1) field campaign, 1419 whole air samples were collected over the Southern Ocean, of which approximately 700 samples were collected in the marine boundary layer (MBL), 300 samples were taken in the free troposphere (FT), and the remainder were collected in the buffer layer (BuL), the layer between the MBL and FT. Concentrations of tetrachloroethene, ethane, ethyne, and propane decayed over the 24 day duration of the intensive portion of the field campaign, which began November 18, 1995. This decline was consistent with what is known about seasonal increase of HO and the seasonal decrease in biomass burning. Using a simple empirical model, the best fit to the observations was obtained when the average [HO] was 6.1 + 0.3 x 105 HO cm -3, and an average [C1] of 720 + 100 C1 cm -3. The corresponding exchange times were 14 + 2 days between the MBL and FT, and 49 +40/-13 days between the MBL in the intensive campaign region and the MBL region to the north (nMBL).

The NOy reactions generally proceed as

XCl(gas) + hv --• C1 + products (2)
where XCI is an easily photolyzed CI species, such as C12, NOC1, C1NO 2, or HOC1, which can accumulate overnight and dissociate within about 1 hour after sunrise [Green, 1972;Finlayson-Pitts, 1983 hydrocarbons not only provide information on the average oxidative capacity of the atmosphere, but also on its seasonality. Seasonal variations of NMHCs and tetrachloroethene in the northern hemisphere (NH) [e.g., Blake and Rowland, 1986;Singh and Zimmerman, 1992;Penkett et al., 1993;Rudolph, 1995;Wang et al., 1995] have been reported but few observations exist for the southern hemisphere (SH) [Blake and Rowland, 1986;Rudolph et al., 1989;Wang et al., 1995]. Maximum background hydrocarbon levels are observed in winter months, with minimum levels occurring during summer. In the NH, longer lived gases tend to reach their minimum concentration later in the solar year compared to gases that react more quickly with HO [e.g., Penkerr et al., 1993: Jobson et al., 1994bGoldstein et al., 1995].
The International Global Atmospheric Chemistry (IGAC) sponsored First Aerosol Characterization Experiment (ACE 1) field campaign intensive period, mid-November to mid-December 1995, was conducted over the pristine Southern Ocean region south of Australia. The primary focus of ACE 1 was to observe the formation and growth of new particles from their precursors and observe their effect on cloud formation and radiative properties ]. Measurements were made at two coastal sites, aboard two research ships, and aboard the National Center for Atmospheric Research (NCAR)

Experiment
More than 2500 whole air samples were collected The fourth GC contained two columns; a 60 m x 0.25 mm CD-B/Cyclodex column coupled to an FID, which separated and detected higher molecular weight NMHCs, and a composite column, which consisted of a 30 m of DB-1 column spliced to 30 m of a DB-5 MS column, which along with an ECD, separated and detected C1 and C2 halocarbons. The detector signals were sent to six Spectra Physics 4400 computing integrators and an IBM 486 personal computer using Labnet software for data capture and storage. This configuration allowed for six simultaneous yet separate chemical analyses from the same sample.
For ACE 1, a 1520 cm 3 (at STP) aliquot of each sample was trapped on a preconcentration loop filled with 3 mm diameter glass beads, immersed in liquid nitrogen. Once the sample was trapped, the preconcentration loop was isolated and warmed to 80øC, then injected and directed by hydrogen carrier gas to the GCs. The split was reproduced with high precision, and 27.0%

Eulerian Observations and Model Description
Meteorological parameters provided by NCAR describing relative humidity, temperature, potential temperature (THETA), and equivalent potential temperature (THETA E) enabled the samples from each flight to be grouped into three vertical layers, the marine boundary layer (MBL), free troposphere (FT), and buffer layer (BuL). Typically, the MBL extended from the surface to about 500-1000 m, while the FT was above an altitude of approximately 3000 m. One example from flight 18 is shown in Figure 1. Figure 2 displays the average concentrations and one standard deviation of the mean error bars for tetrachloroethene, ethane, ethyne, and propane for each of the local flights. Best fit curves to the average temporal mixing ratio data were calculated, employing a function with two exponential terms, in order to approximately describe atmospheric mixing and photochemistry. The average FT mixing ratios are plotted in As seen in Figure 2, the best fit curve for tetrachloroethene and ethane appear to be at the inflection of their seasonal cycles, and closer to their maximums than ethyne and especially propane, which are closer to their minimums as indicated by their decreasing negative slopes. the exchange times between the MBL and the FT ('rexl) , as well as between the MBL and the nMBL, 'rex 2. The concentrations in the buffer layer (BuL) were very similar to those measured in the MBL, and mixing between the MBL and the BuL mainly recycles air between these two layers. Thus this layer acts to "buffer" dilution by impeding direct mixing with the FT as described by Wingenter et al. [1996]. Therefore this layer was not included in the following calculations. A detailed A   Mixing ratios are in pptv.

Discussion and Comparison of ACE 1 Eulerian Results
Four day back trajectory analysis [Whittlestone et al., 1998] (for an overview of the meteorology, see also Hainsworth et al. [1998]) shows that the air parcels reaching the intensive sampling region originated between 40øS and 60øS latitude, with a mostly westerly flow. This indicates that the air masses were representative of the mid-latitude/subpolar region. Potential interference with our results could occur if direct emissions of tetrachloroethene, ethane, ethyne, and propane were introduced upstream of the ACE 1 intensive region, and remained in the MBL. However, the homogeneous air masses encountered indicate a lack of recent emissions in the region (see Figure 4 and Table 3   bThe average mixing ratios were determined from the best fit curves in Figures 2-5

Comparison of the Oxidative Capacity of HO and CI
Chemical loss by C1 in the MBL during the springtime Southern Ocean is significant for tetrachloroethene, ethane, propane, and to a lesser extent ethyne (Table 2). Chemical loss by C1 and HO at the concentrations determined in this work are also compared in Table 2 for some stratospheric ozone-depleting substances such as methyl bromide and methyl chloroform, the greenhouse gas methane, and DMS, the main precursor to new sulfate particle growth in the clean MBL [Clarke et al., 1998b]. At the average C1 levels encountered during ACE 1, C1 oxidation has little impact on methyl bromide, methyl chloroform, or methane. However,

Model Limitations
This three.-box model has certain limitations, namely, the concentrations of the three hydrocarbons and tetrachloroethene to the south of the intensively studied MBL were not measured befoxe and after the project, so that the effects on the ACE 1 intensive MBL from the southern area cannot be directly ascertained. However, when the temporal decrease is removed from the four gases in question and their normalized concentrations are plotted versus latitude, as shown, in Figure 4, along with the best linear fit, very shallow latitudinal gradients are apparent with R 2 of 0.01 1 at best to 0.00012 (see Table 3 for results). This, and the fact that surface winds were predominantly zonal, leads to the conclusion that merJchonal mixing was not a major factor in the ACE 1 study region. This point was examined further by not including the nMBL in a model simulation. Although S increased by a factor of 10, the values for HO, C1, and vertical mixing remained unchanged within the uncertmntles cited above. Therefore, if the southern region influence was similar in magnitude to the nMBL, our estimates of HO, C1, and vertical mixing are still accurate.
Another limitation of this model is that four parameters are being predicted from the time series of only four gases. Unfortunately, shorter lived gases such as butanes or larger hydrocarbons were often times below detection limit and therefore, unsuitable for the analysis in this paper. Gases such as CFCs or HCFCs are too long lived in comparison to the timescale of this experiment to be useful. Therefore, other possible factors, which are unknown, cannot be detected by their impact on the time series of any other additional gases. However, the close agreement between measured and inferred HO gives confidence to the model. The close agreement between the vertical mixing timescale estimated by the three box model and the methyl nitrate vertical profile also increases the confidence in the model. In fact, the very good agreement between the measured and modeled time series for tetrachloroethene, ethane, ethyne and propane leads to the conclusion that no other parameters are likely significant in contributing to the atmospheric MBL burden of these gases in the ACE 1 intensive study region.

Conclusion
The average hydroxyl estimate of 6.1 + 0.3 x 105 HO cm-3 for the MBL over the Southern Ocean during the springtime ACE 1 project agrees well with the measured diurnal average. The corresponding C1 concentration estimate in this pristine region of 720 + 100 C1 cm-3 may help elucidate the mechanism(s) responsible for C1 production. Although C1 concentrations were 3 orders of magnitude less than HO, C1 is a significant oxidizer for some trace gases in the MBL because of its greater reactivity with these compounds. Transport times from the MBL to the FT and from the nMBL to the MBL in the intensive study region were 14 +_ 2 days and 49 + 40/-13, respectively. The methyl nitrate vertical gradient was employed for validation of the MBL/FT exchange time calculated using the three-box model. Results from this study may help constrain 2-D and 3-D model values of HO and C1 concentrations as well as transport time scales in this region.