Distribution of halon-1211 in the upper troposphere and lower stratosphere and the 1994 total bromine budget

. We report here on the details of the first, in situ, real-time measurements of H-1211 (CBrC1F2) and sulfur hexafluoride (SF6) mixing ratios in the stratosphere up to 20 km. Stratospheric air was analyzed for these gases and others with a new gas chromatograph, flown aboard a National Aeronautics and Space Administration ER-2 aircraft as part of the Airborne Southern Hemisphere Ozone Experiment/Measurements for Assessing the Effects of Stratospheric Aircraft mission conducted in 1994. The mixing ratio of $F6, with its nearly linear increase in the troposphere, was used to estimate the mean age of stratospheric air parcels along the ER-2 flight path. Measurements of H-1211 and mean age estimates were then combined with simultaneous measurements of CFC-11 (CC13F), measurements of brominated compounds in stratospheric whole air samples, and records of tropospheric organic bromine mixing ratios to calculate the dry mixing ratio of total bromine in the lower stratosphere and its partitioning between organic and inorganic forms. We estimate that the organic bromine-containing species were almost completely photolyzed to inorganic species in the oldest air parcels sampled. Our results for inorganic bromine are consistent with those obtained from a photochemical, steady state model for stratospheric air parcels with CFC-11 mixing ratios greater than 150 ppt. For stratospheric air parcels with CFC-11 mixing ratios less than 50 ppt (mean age >5 years) we calculate inorganic bromine mixing ratios that are approximately 20% less than the photochemical, steady state model. There is a 20% reduction in calculated ozone loss resulting from bromine chemistry in old air relative to some previous estimates as a result of the lower bromine levels. method is based on simultaneous measurements of solar radiation profiles of [CH2Br 2] and [CH2BrC1] in the troposphere have been measured by the authors (E.L.A and S.M.S), and they exhibit substantial variability. With the current measurements we estimate that tropospheric degradation of these two species does not exceed 50%. Profiles of [CH3Br] in the troposphere (E.L.A. and S.M.S.) also exhibit some variability, and from these profiles we estimate that the mixing ratio of CH3Br at the tropopause differs from the global surface average by less than 10%. Our estimates of BrTota I would be reduced by a little more than 2 ppt if significant degradation is occurring in the troposphere. The measurements of H-1211 and estimates of the remaining species difference between total bromine and organic bromine.


Introduction
Concern over the contribution of bromine to stratospheric ozone loss in polar regions [McElroy et al., 1986;Salawitch et al., 1988;Anderson et al., 1989] and in midlatitudes [Yung et al., 1980;Garcia and Solomon, 1994;Wennberg et al., 1994] has resulted in international regulation of the production of halons and methyl bromide (CH3Br) [United Nations Environmental Program, 1995].
The concentration of bromine in the stratosphere is about 2 orders of magnitude smaller than the concentration of chlorine, but bromine is 40-100 times more efficient at destroying ozone than is chlorine [Garcia and Solomon, 1994;Solomon et al., 1995]. Bromine accounts for 20-40% of the annual loss of stratospheric ozone in polar regions [Anderson et al., 1989;Salawitch et al., 1993] and plays a significant role in midlatitude ozone loss at altitudes below 20 km [Garcia and Solomon, 1994;Lary, 1996].
Organic compounds containing bromine are released at the Earth's surface (Table 1), mix rapidly in the troposphere, and are advected into the stratosphere where photolysis and reaction with OH lead to the formation of inorganic bromine species. It is the inorganic species that participate in the catalytic destruction of ozone. A recent modeling study found that at an altitude of 20 km in the midlatitudes, the inorganic bromine species, Br 2, BrO, BrONO, BrC1, BrONO2, and HOBr, have lifetimes in sunlight of less than 1 hour [Lary, 1996]. The model results indicate that the longest-lived inorganic bromine species is HBr with a lifetime of 1 day. The model results also suggest that only 1% of the inorganic bromine is in the form of HBr and that approximately one half of inorganic bromine is present as BrO in sunlight in the lower stratosphere [Lary, 1996]. Although inorganic bromine is involved in catalytic ozone destruction in the absence of reactive 1513  nitrogen, hydrogen, and chlorine species, bromine also participates in "mixed-family" cycles of catalytic ozone destruction. In addition to producing ozone loss, the "mixedfamily" reactions influence the partitioning within the bromine, chlorine, nitrogen, and hydrogen families [Dandin and McConnell, 1995;Lary et al., 1996] which makes the contribution of bromine to ozone loss depend strongly upon local conditions [Garcia and Solomon, 1994;Danilin et al. , 1996]. It is important to note that our understanding of bromine chemistry in the stratosphere is not yet complete as evidenced by the significant changes and additions for bromine chemistry between the JPL-94 and JPL-97 recommendations [DeMote et al., 1994, 19971. Halons, which are purely anthropogenic compounds used primarily as fire suppressants, and CH3Br, a naturally occurring compound that is also manufactured and used as a fumigant, are the major sources of stratospheric bromine ( Table 2). The halons contribute about one third of the total bromine enterine the stratosphere. The sale and manufacture of halons was banned in developed countries on January 1, 1994, and is scheduled to be banned in developing countries by 2010 [United Nations Environmental Program, 1995]. Halons have been stockpiled and their use limited to critical applications such as archive protection and passenger protection aboard airplanes, ships, and military vehicles. Although the tropospheric mixing ratios of the two most abundant halons, H-1211 and H-1301 (CBrF3), continue to rise, their combined tropospheric growth rate has recently slowed [Butler et al., 1992;Montzka et al., 1996, Butler et al., 1998]. The largest single source of stratospheric bromine, CH3Br, has a tropospheric mixing ratio of 9.8 +_ 0.6 ppt [Lobeft et We have assumed that the mixing ratios of the naturally occurring species were constant from 1988 to 1995. We allowed for changes in the mixing ratios by including uncertainty in the linear growth coefficient. The time series of halon mixing ratios at the Earth's surface come from flask measurements collected at NOAA/CMDL's seven baseline monitoring stations. The trends for H-1211 and H-1301 are based on measurements that began in mid-1989. The trend for H-2402 is based on measurements that began in 1995 and archived air samples. The mixing ratio as a function of time, IX(t)], is given by the equation, IX(t)] = a + b(t -1994.875) + c(t -1994.875) 2. We have chosen mid-October 1994 as our reference time because this coincides with the in situ stratospheric measurements taken during the ASHOE/MAESA mission. With this convention, the constant term, a, is the surface mixing ratio for mid-October 1994. The uncertainties are I standard deviation values. al., 1995; Penkett et al., 1995]. Methyl bromide is produced and destroyed by a variety of physical and biological processes on land and in the ocean. The release of CH3Br to the atmosphere during fumigation is the largest anthropogenic source of CH3Br; smaller contributions arise from biomass burning and use of leaded fuel additives. Naturally occurring CH3Br is believed to account for 60-80% of the total atmospheric burden of methyl bromide [Butler and Rodriguez, 1996;?enkett et al., 1995]. Additional source gases include the relatively short-lived species, dibromomethane (CH2Br2) and bromochloromethane (CH2BrC1), which have atmospheric lifetimes on the order of a year and tropospheric mixing ratios of 1.1 and 0.14 ppt (Tables 1 and 2). Several other naturally occurring, organic, brominated compounds have been measured in the troposphere, but they are short-lived in the atmosphere and are believed to be insignificant contributors to stratospheric bromine ( Table 1).
The amount of bromine in the stratosphere and its partitioning among inorganic and organic species are crucial inputs for atmospheric models that attempt to simulate the distribution of ozone and evaluate the impact of HOx, NOx, and halogen chemistry on ozone destruction [Salawitch et al., 1993' Wennberg et al., 1994' Gao et al., 1997. Currently, in situ techniques measure a limited number of brominated compounds from which total bromine is estimated. We calculate the total bromine in stratospheric air parcels by first determining the time series of bromine input to the stratosphere. The amount of bromine in an isolated, stratospheric air parcel remains constant at the value it had upon entry to the stratosphere, though photolysis and reaction with OH convert the organic source compounds to inorganic species as the parcel passes through the stratosphere. For a real air parcel, subject to large-and smallscale mixing processes in the atmosphere, the amount of bromine will reflect contributions from air parcels that entered the stratosphere at different times. However, there remains a correlation between the total bromine and the mean age of the air parcel. The mean age of stratospheric air is determined from the measured mixing ratio of an age tracer such as SF 6. The amount of bromine present in organic forms is either measured directly or estimated from tracer-tracer correlations.
The amount of inorganic bromine in an air parcel is then determined by subtracting the organic bromine from the total bromine. This method is based on measured mixing ratios of organic, brominated source compounds in both the troposphere and stratosphere and will therefore be referred to as the organic bromine method. An independent method for determining the amount of inorganic bromine in a stratospheric air parcel makes use of the rapid photochemical cycling among various inorganic, bromine-containing species in the sunlit stratosphere. This method is based on simultaneous measurements of solar radiation and mixing ratios of BrO and other radicals and will be referred to as the inorganic bromine method. Both methods were used to calculate the partitioning of bromine in the lower stratosphere from measurements taken during the 1994 Airborne Southern  [Elkins et al., 1993[Elkins et al., , 1996b. Most of the measurements were obtained in the stratosphere at ER-2 cruise altitudes of 15-20 km, but a few tropospheric measurements were obtained upon ascent and descent of the aircraft (Figure 1). Measurements were taken in October and November 1994 at latitudes ranging from 70øS to 60øN.
All prior measurements of organic, bromine-containing compounds in the stratosphere had been made by collecting whole air samples from aircraft or balloon platforms and storing them for analysis with ground-based instruments. In this work we incorporated measurements of brominated compounds in samples collected by researchers   (1) By convention, Bry designates bromine residing in inorganic species, CBr¾ designates the bromine contained in organic bromine spedies, and the total amount of bromine is given by BrTota 1. The use of square brackets indicates a mixing ratio rather than a concentration throughout this paper. The only significant source of bromine in the stratosphere is assumed to be organic bromine emitted at the Earth's surface (Tables 1 and 2). We also assume that only those species with lifetimes in the troposphere that are long enough to allow them to be advected into the stratosphere (lifetime > 0.4 years) contribute to stratospheric bromine. As a result, we assume in our calculation that the stratospheric input function, BrTotal, is the sum of the bromine contained in the six species, CH3Br, H-1211, H-1301, H-2402, CH2Br 2, and CH2BrC1 (Table 2). First, to produce a sum, the time series of the mixing ratios of these six compounds at the Earth's surface was compiled. The use of surface measurements rather than measurements at the tropical tropopause sets an upper limit on how much bromine can be carried into the stratosphere by these species. Second, the mean age of air along the ER-2 flight path was estimated from measurements of SF 6, a longlived, anthropogenic gas with a growth rate of approximately 6.7% There are few published tropospheric measurements of CH2Br 2 and CH2BrC1. We assigned 1994 tropospheric mixing ratios of 1.1 _+0.2 ppt for CH2Br2 and 0.14 _+0.03 ppt for CH2BrC1, based on NCAR analysis of air samples collected during the May 1995 STRAT campaign. There are no trend data available for the tropospheric mixing ratios of these two naturally produced compounds, so we assumed that their tropospheric mixing ratios were constant over the 6 year period of interest. To compensate for this assumption, we adopted 1 standard deviation (1 s.d.) uncertainties of _+0.1 ppt/year and _+0.01 ppt/year in the trends of CH2Br 2 and CH2BrC1. This assumption encompasses extremes as the 2 standard deviation uncertainty includes the possibility that the tropospheric mixing ratios of these two compounds grew linearly from zero to their late 1994 tropospheric values over a period of 6 years.
The stratospheric input function for organic bromine is estimated as the sum of the bromine contained in six long-lived, organic bromine species (Table 2) where the mixing ratio for each species is measured at the tropical tropopause. The time series for each source compound given in Table 2  where all mixing ratios are in ppt. The halons are assumed to account for 100% of the increase in tropospheric mixing ratios of compounds that contribute to stratospheric bromine (Table 2, Figure 2). The uncertainty in [BrTota •] was calculated from the uncertainties in the late 1994, tropospheric measurements and the uncertainties in the tropospheric trends of each species (Tables 2,   3, and 4).

Mean Age of Air
Species with no appreciable sources or sinks in the stratosphere that also have linearly increasing tropospheric mixing ratios can be used to determine the mean age of air in stratospheric air parcels [Hall and Plumb, 1994]. Use of a conservative tracer allows one to evaluate the mean age of an air parcel by comparing the measured mixing ratio in the stratospheric parcel to a value in a time series of tropospheric mixing ratios. The mean age of air can be estimated from the mixing ratios of several different tracers, including CO 2, CFC-115, and SF6 [Woodbridge et al., 1995;Daniel et al., 1996;Harnisch et al., 1996]. The relationship between the mean age, F, the stratospheric mixing ratio, [X], and the mixing ratio at the tropopause, IX]o, is given by equation (4).

[X](x,t) = [X]o (t-F(x)) (4)
The measurement in the stratosphere occurs at a point in space and time designated by (x,t). The mean age is referenced from the time an air parcel crosses the tropical tropopause. The mean age of air cannot be obtained directly from measured mixing ratios of an age tracer if the time series of tropospheric mixing ratios for the age tracer is not linear. A complete discussion of the method for deriving the mean age of stratospheric air parcels from measurements of SF 6 can be found elsewhere . For a species with a quadratic trend, equation (4) becomes Stratospheric measurements of the compounds that contribute to stratospheric bromine from four research groups were used to obtain a correlation between the mixing ratios of each brominated compound and the mixing ratio of CFC-11. To compensate for the different calibration scales used by each research group, the data from different research groups was normalized so that each species had a normalized value of one at the tropopause.
The presence of a quadratic term in the time series produces a deviation away from the strictly linear case that is proportional to the quadratic coefficient and the mean age of the air. Sulfur hexafluoride is a purely anthropogenic compound that is used widely as an electrical insulator in transformers and high voltage switching devices. Very small amounts have also been released during tracer studies. It has no known tropospheric sinks and a stratospheric lifetime of approximately 3200 years [Ravishankara et al., 1993]. The recent trend in global, surface mixing ratios (in ppt) of SF 6 is slightly nonlinear [Geller et al., 1997].
[SF6]surf(t ) = 3.4361 + 0.2376(t -1996) + 0.0049(t -1996) 2 (7) Equation (7) is valid for 1987 < t (years) < 1997. The mixing ratio of SF 6 at the tropical tropopause will be less than that at the surface due to the increasing surface mixing ratios, the finite Equation (11) is a second-order polynomial fit to all data, including data taken in the tropics, with each data point weighed equally. This correlation, which forms a snapshot of the relationship between SF 6 and CFC-11 at the end of 1994, will change with changes in the tropospheric mixing ratios of these two compounds.

Estimates for Brominated Source Species Not Measured During ASHOE/MAESA
Correla•tions of brominated species with CFC-11 were compiled from analyses of whole air samples collected by the UC-Irvine (DC-8, 1992), UEA/KFA (four balloon flights from March 1993 to July 1994), and NCAR (ER-2, 1995) samplers. As previously mentioned, the correlation between two species will depend in part upon the tropospheric mixing ratio of each species. The three groups did not measure the same tropospheric mixing ratios of the various brominated compounds (Table 4), and the spread in the measurements exceeds the observed global trend for each species over the measurement time period. Thus we would not expect correlations obtained by different groups to be in good agreement. The spread in measurements may arise from real atmospheric variability, a systematic error in the measurements, or simply a scalar calibration error. Each laboratory used an independent calibration scale which could account for the spread in the measurements. Scalar calibration differences can be resolved by applying a species-and labdependent normalization factor to the original measurements. We normalized the data from different research groups so that each species had a value of one at the tropopause. The normalized correlations obtained by the various groups were in much better agreement than the unnormalized correlations. The scatter in the normalized correlations is the result of measurement uncertainty and atmospheric variability (Plate 1). There are not enough data to extract the seasonal and spatial dependencies of the atmospheric variability and this prevents a more precise parameterization of the correlations. For this reason we combined all data to obtain an average correlation for each

species. A simple parameterization of the relationship between the mixing ratio of a species at the tropopause, [X] o, and the mixing ratio of the species in the stratosphere, [X], is given by equation (12). [X] = [X] o exp(-t/x x)
The quantity, 1/Xx, is an effective stratospheric loss rate. For the species being considered here, the dominant loss mechanism is photolysis. The correlation between two species with stratospheric mixing ratios that are well represented by equation (12) is a power law relationship with a power coefficient equal to the ratio of the loss rates of the two species. The correlations between the brominated species and CFC-11 are adequately described by a simple power law (equation (13), Plate 1),

[X]/[X]o = ([CFC-11]/[CFC-11 ]o) l/d (13)
where the mixing ratio of each brominated species is represented by [X], the subscript, o, designates the mid-October 1994 mixing ratio at the tropopause, and (l/d) is the power coefficient (Table  4, Plate 1). While not rigorously accurate, the value of d can be used to provide an estimate of the loss rate of each species relative to that of CFC-11. These correlations were obtained by weighing each data point equally, including data taken in the tropics. For air parcels sampled along the flight track of the ER-

2, [CBry] was calculated from the direct measurement of H-1211
and estimates of the mixing ratios of the five unmeasured organic bromine species from these correlations with CFC-11. The uncertainty in [CBry] was calculated from the uncertainty in the measurements of the late 1994 tropospheric mixing ratios (Table  2) and the uncertainty in the value of d for each species (Table 4).

Organic bromine can also be estimated solely from measurements of [CFC-11 ] if the correlation between [H-1211 ]
and [CFC-11 ] is used rather than measured H-1211 mixing ratios. The correlations discussed above were based upon the mixing ratio of CFC-11 rather than that of nitrous oxide (N20) because the lifetimes of the brominated source compounds are closer to the lifetime of CFC-11 than to the lifetime of N20. However, the same methods could have been applied using the measurements of N20, which were obtained every 6 min by GC-ECD and at 1 Hz using a tunable laser absorption technique [Podolske and Loewenstein, 1993]  and is valid for N20 mixing ratios between 130 and 310 ppb. The mixing ratio of CFC-11 is in ppt and the mixing ratio of N20 is in parts per billion (ppb) for equations (15) and (16).

Results and Discussion
The observed distribution of H-1211 mixing ratios measured during ASHOE/MAESA was consistent with that expected for a species with an increasing tropospheric source and a stratospheric sink subject to Brewer-Dobson circulation (Figure 4). Air parcels sampled on ascent and descent (data points below 15 km) from (21øN, 158øW), Nadi, Fiji (18øS, 175øE), and Christchurch, New Zealand (43øS, 172 ø E) show values consistent with late 1994 surface mixing ratios. At a given altitude, mixing ratios of H-1211 in the tropics were larger than those at higher latitudes, consistent with the upward advection of tropospheric air that contained high mixing ratios of H-1211 into the tropical stratosphere. The lowest H-1211 mixing ratios were observed in air parcels located at high altitudes. On the flight of October 13, 1994, low H-1211 mixing ratios were observed at 16 km when the ER-2 executed a dive near 68øS. This is consistent with downward advection of air in polar regions.

Moffett Field, California (37øN, 122øW), Barber's Point, Hawaii
A similar picture emerged from the SF 6 measurements (data not shown). The highest values of [SF6] were observed in the tropics with a trend toward lower values with increasing altitude and latitude. The mean age of air calculated from the measurements of SF 6 (equation (10)) ranged from 0 to 6 years [Waugh et al., 1997;Volk et al., 1997] (Figure 5). At high northern and southern latitudes, air parcels were sampled below 18 km with mean ages greater than 4 years, which suggests vertical descent of stratospheric air in these regions. There was only one air sample with a calculated age in excess of 6 years. In general, the oldest air parcels sampled were close to the polar region and had mean ages of 5 to 5.5 years.  ........... ............. ........................................ o.0'o ......  ".. : .................... ' .............. (Table 5). On the basis of the tropospheric mixing ratios and the lifetimes of the organic species that contribute to halogens in the stratosphere, we expect the conversion of organic bromine species to inorganic bromine species to be more complete than the analogous conversion of chlorine species in any given air parcel. Consistent with this, we calculate that the conversion of CBry to Bry of 97.5% is more complete than the conversion of CCI•, to Cly in air parcels with similar ages observed during AASE-If (Table 5).

We calculated [BrTotal] , [CBry], and [Brv] in air parcels
The results in Figure 6 and the accompanying equations and tables are representative of the lower stratosphere in late 1994. Since the results are expressed in terms of absolute mixing ratios, anticipated changes in tropospheric abundances of CFC-11, the halons, and CH3Br will necessarily lead to modified relationships in the future. Because regulation of the halons will substantially reduce their tropospheric mixing ratios, the minimum [Bvrotal] in the stratosphere is expected to be set in the future by the tropospheric mixing ratios of naturally produced, brominated compounds. We estimate this value to be approximately 9.2 ppt under the assumptions that 70% of the CH3Br burden [Butler and Rodriguez, 1996 Although the values of [Bry] calculated with the inorganic bromine method agree within combined uncertainties with those calculated from the organic bromine method (Figure 7), the value of the intercomparison is lessened by the greater uncertainties inherent in the inorganic bromine method. The organic bromine method results in significantly less uncertainty. The values of [Bry] calculated with the inorganic bromine method are at the upper limit of what we would calculate from the organic bromine method. We note, however, that there is a potential systematic error in our initial assumption regarding the source of stratospheric bromine (Table 2, equation (2)). If additional sources of stratospheric bromine are significant, our estimate of [BrTotal], and thus [Bry], will be too low.
The organic bromine method of calculating the BrTota I in the stratosphere works best for long-lived source gases which have negligible sinks in the troposphere. For these species, input to the stratosphere can be well parameterized by the global surface mean. Short-lived species which can exhibit large deviations The mixing ratio of N20 associated with the calculated fractions of inorganic chlorine and bromine has been included as a reference point for the mean age of the sampled air independent of the method used to determine the mean age of the air. away from a global average are not as well represented in our analysis. One such compound is bromoform (CHBr3) which is naturally produced and has mixing ratios at the Earth's surface between 0 and 10 ppt [Penkett et al., 1985;Cicerone et al., 1988;Atlas et al., 1992]. Several measurements of CHBr 3 mixing ratios in excess of 20 pptv were obtained by Class et al. [ 1986] in a 1985 survey of halomethanes in marine air. Decomposition of 1 ppt of CHBr 3 in the stratosphere would increase [BrTota 1] by 3 ppt, which is more than 10% of the current input to the stratosphere. The lifetime of CHBr3 is estimated to be of the order of weeks [Penkett et al., 1985], so CHBr 3 was not included as a source of stratospheric bromine in our estimate of [BrTotal]. Bromoform mixing ratios of 0.1 ppt have recently been measured (E. Atlas, unpublished data, 1996) in whole air samples collected in the upper troposphere, but no bromoform has been measured in the stratosphere. Recent results from the AER, 2-D chemical transport model suggest that the organic species with lifetimes less than a month may be contributing appreciable amounts of bromine to the stratosphere [Ko et al., 1997]. Determining the amount of CHBr 3 and other short-lived, organic compounds of bromine, or inorganic bromine from their decomposition, entering the stratosphere remains an outstanding issue in the bromine budget of the lower atmosphere.

Conclusion
Stratospheric measurements of SF 6 were combined with the tropospheric time series of mixing ratios for six organic bromine species to calculate the amount of total bromine in stratospheric air parcels sampled at latitudes between 70øS and 60øN to altitudes of 20 km during late 1994. The oldest air parcel sampled had a mean age of 6 years and contained an estimated 16.4 _+2 ppt of bromine. Air parcels that had recently entered the stratosphere contained an estimated 18.2 +2 ppt of bromine. Total organic bromine in each air parcel was calculated from measurements of H-1211 and estimates of the remaining species based upon correlation with CFC-11 obtained from independent stratospheric data sets. The good agreement found between the observed correlations and those generated by a 2-D chemical transport model suggests that the conversion of organic to inorganic bromine in the stratosphere is reasonably well understood. Inorganic bromine was calculatecl from the difference between total bromine and organic bromine.
The method described in this paper yields estimates of [BrTotal] and [Bry] in stratospheric air parcels with smaller uncertainty than other methods. The uncertainty could be reduced with more precise, global monitoring of brominated source gases at the Earth's surface and an expanded measurement program for organic and inorganic bromine compounds in the lower atmosphere. Altitude profiles of short-lived organic bromine species, CHBr 3 in particular, from the Earth's surface to the tropopause would be helpful in addressing this issue. Our method also presumes that strong, localized convection contributes negligible amounts of bromine to the stratosphere. Firmer conclusions are possible only after the contributions of short-lived organic bromine species to stratospheric bromine become better quantified.
Even though the organic and inorganic methods described here for calculating Br v agree within their respective uncertainties, further work should be pursued to reduce the larger uncertainties of the inorganic method, primarily through the improved overall accuracy of BrO measurements.