A Net Sink For Atmospheric CH3Br In The East Pacific Ocean

A Net Sink For Atmospheric CH₃Br In The East Pacific Ocean

Jürgen M. Lobert, ^‡† James H. Butler, ^‡ Stephen A. Montzka, ^‡
Laurie S. Geller, ^‡† Richard C. Myers, ^‡ James W. Elkins ^‡

Surface waters along a cruise track in the East Pacific ocean were undersaturated in methyl bromide (CH₃Br) in most areas except for coastal and upwelling regions, with saturation anomalies ranging from +100 percent in coastal waters to -50 percent in open ocean areas, representing a regionally weighted mean of 16 (-13 to -20) percent. The partial lifetime of atmospheric CH₃Br with respect to calculated oceanic degradation along this cruise track is 3.0 (2.9 to 3.6) years. The global, mean dry mole fraction of CH₃Br in the atmosphere was 9.8 ą 0.6 parts per trillion, with an interhemispheric ratio of 1.31ą 0.08. These data indicate that ~8 percent (0.2 parts per trillion) of the observed interhemispheric difference in atmospheric CH₃Br could be attributed to an uneven global distribution of oceanic sources and sinks.

Paper published in Science Vol. 267, 1002-1005 17-February-1995.

Methyl bromide (CH₃Br) has received considerable attention recently (1-11), in particular because of its high ozone depletion potential (5, 6). The tropospheric mixing ratio of organic bromine is believed to be ~20 parts per trillion (ppt), to which CH₃Br contributes ~50% (5, 7). However, the atmospheric CH₃Br budget is uncertain at this time, particularly with regard to oceanic contributions (1, 2, 3, 10).

Annual fluxes of CH₃Br to the atmosphere could range from 76 to 198 Gg (1 Gg = 10⁹ g), based on reported atmospheric mixing ratios (2) and lifetimes (12, 13). Emissions from industrial activities (2 to 10%), soil fumigation (20 to 60%), and biomass burning (10 to 50%) have been estimated to contribute 33 to 74% to the total flux of CH₃Br to the atmosphere (5, 8, 9); the remainder of global emissions has historically been attributed to the oceans. Singh et al. (10) found saturation anomalies of 180 to 240% for the East Pacific and Khalil et al. (3) found 40 to 80% for the Pacific Ocean. The reasons for the discrepancy between these studies are unknown, but it could be caused by sampling and analytical differences or by geographic differences in the distribution of dissolved CH₃Br. Singh et al. (10) measured CH₃Br mainly in productive nearshore waters and areas close to the upwelling regions of the Americas, whereas Khalil et al. (3) derived their estimate mainly from the open ocean. Singh et al. compared in situ atmospheric partial pressures with surface water concentrations, which required solubility data to obtain saturation anomalies, whereas Khalil et al. collected air in stainless steel flasks and directly compared partial pressures of CH₃Br in the atmosphere with those above equilibrated surface water.

For the National Oceanic and Atmospheric Administration (NOAA)-Climate Monitoring and Diagnostics Laboratory Bromine Latitudinal Air/Sea Transect 1994 (BLAST 94) expedition, a transit leg of the World Ocean Circulation Experiment Line P18 was selected because it covered coastal waters, central oceanic gyres, and regions of divergence and upwelling in both hemispheres, as well as two different seasons. We made measurements aboard ship to eliminate artifacts associated with storing samples and partial pressures were measured directly to remove the dependence of saturation anomalies upon currently uncertain solubility data. A mass spectrometer (MS) was selected over an electron capture detector (ECD) for measurement of CH₃Br to circumvent potential interferences common to the use of ECDs.

Fig. 1. Cruise track of NOAA Ship Discoverer, January-February 1994. Horizontal dashed lines and labels indicate oceanic regions used for computation of results: region 1, open ocean (winter); region 2, coastal waters (winter); region 3, divergence between North Equatorial Current (NEC) and North Equatorial Counter Current (NECC); region 4, equatorial divergence (upwelling); region 5, open ocean (summer); region 6, coastal and nearshore waters (summer); and region 7, inland passage.

Research was conducted aboard the NOAA Ship Discoverer between 26 January and 18 February, 1994. The 13600 km cruise track led from 48°N to 54°S in the East Pacific ocean (Fig. 1.). Ambient air was sampled at the most forward point of the ship ~15 m above the ocean surface. For the partitioning of gases in surface seawater, we used an acrylic equilibrator (14, 15). The approximate residence time of air and equilibrator headspace in the sampling setup was ~1 min, the average sample spacing was ~30 km (one sample per hour). Instrumental precision (1 sigma) for CH₃Br was better than 2%; the detection limit (3 sigma) was ~0.5 ppt of CH₃Br (16). Measurements were standardized with frequent injections of natural air, calibrated with gravimetric standards prepared in our laboratory.

Several tests were conducted at sea to ensure that the data were not marred by analytical or sampling artifacts. Although CH₃Br was usually detected as [CH₃⁸¹Br]⁺ (mass to charge ratio m/z⁺=96), it occasionally was monitored as [CH₃⁷⁹Br]⁺ (m/z⁺=94) to assure that mass fractionation or an interference by other compounds was not affecting the results. No peaks with ion 94 or 96 eluted near CH₃Br. No blank signals were detected when purified, halocarbon-free air was periodically injected. Finally, the equilibrator was tested for potential internal loss or production of analyte gases, and an alternative sampling scheme, which allowed for sampling without any devices other than valves in the sample stream, was used on occasion. All of these tests showed no analytical or sampling interferences that would significantly affect the determination of CH₃Br saturation anomalies.

Flasks of ambient air and equilibrator headspace were filled every few degrees of latitude. All flasks were analyzed with separate gas chromatography-mass spectroscopy (GC/MS) and GC-ECD instruments over a period of several months after the cruise. Most of those results agreed with shipboard data, although some flasks showed significant differences owing to storage effects. We confirmed the undersaturations determined by the two GC/MS systems with a GC/ECD system equipped with a Poraplot Q column. The results of GC/ECD analysis made with a DB-1 column, however, disagreed and, in fact, showed supersaturations where there were undersaturations according to three other analytical systems.

The smoothed, atmospheric mole fraction obtained from shipboard measurements of CH₃Br ranged from 10 to 12 ppt in the mid- to high latitudes of the Northern Hemisphere (NH) to 7 to 9 ppt in the higher latitudes of the Southern Hemisphere (SH) (Fig. 2A), with mean, hemispheric mole fractions of 11.10.6 ppt (NH) and 8.50.6 ppt (SH) (17). The global, mean mole fraction of CH₃Br was 9.8ą0.6 ppt. Whether the break between hemispheres was selected at the equator (0o) or at the interhemispheric tropical convergence zone (ITCZ; 5°N), the ratio of total CH₃Br between hemispheres (NH/SH) remained at 1.31 (18).

     Click on this image to display the full figure incl legend (75 kb).

Fig. 2: Data on CH₃Br in the air and water of the East Pacific Ocean. Dry mole fraction of atmospheric CH₃Br, (B) its partial pressure in the surface water, and (C) the saturation anomaly (shaded area) and net saturation anomaly (line and markers). Solid lines are LOESS fits (23); the vertical, dashed lines and labels indicate the oceanic regions as described in Fig. 1.

The partial pressure profile of surface water (Fig. 2B) differs distinctly from the atmospheric profile and even more so from data reported by other investigators (3, 10). The partial pressure of CH₃Br in the surface water was generally lower than that in the atmosphere. Further, dissolved CH₃Br did not correlate with dissolved CH₃Cl or CH₃I at a 90% confidence level (correlation coefficient r = 0.061 and 0.196, n = 249), as has been suggested in the past (2, 10); CH₃Cl and CH₃I were highly supersaturated along this cruise track and thus a correlation is not useful for directly predicting oceanic CH₃Br.

The saturation anomaly (19) of CH₃Br appears to be driven mainly by variations in its surface water partial pressure (Fig. 2, B and C). Moderate supersaturations were found off the coast of California and in upwelling regions at 11°N and the equator (1°S), as well as in the coastal inland passage of Chile south of 41°S. Supersaturations as high as 100% were found offshore of Chile (38°S). Most important, however, CH₃Br was undersaturated in the open ocean in both hemispheres. Table 1. Observed values for CH₃Br in three main marine regions. SST, average sea surface temperature as measured during the cruise; V, average true wind speed as observed during the cruise; dCH₃Br, mean saturation anomaly of CH₃Br (19); dCFC-11, mean saturation anomaly of CFC-11 (chlorofluorocarbon-11)(19). Region SST (°C) V (m/s) dCH₃Br (%) dCFC11 (%) Open ocean {1, 5} 21.9 7.1 -21.8 2.5 Coastal {2, 6} 17.6 7.2 40.3 2.9 Upwelling {3, 4} 26.1 6.3 1.4 0.6

To estimate the global significance of these results, mean saturation anomalies were calculated for separate physiographic areas, and our observations within each area were assumed representative for the entire ocean. These divisions were based on oceanic upwelling and current patterns, surface water data on nitrous oxide (20), and abrupt changes in the saturation anomalies of CH₃Br. Data within areas 1 to 6 (Figs. 1 and 2C) were combined into three general regions to obtain means for the open ocean, current divergences, and coastal waters (Tables 1 and 2). Saturation anomalies in the inland passage (region 7) were around +11% and varied greatly, apparently with freshwater inflow. Thus, these anomalies are probably not globally representative for any large area of the ocean and were disregarded in flux computations.

The global mean net saturation anomaly, derived from weighting the three general regions according to their presumed oceanic coverage, was 15.7% (-13 to -20%) (Table 2) (21). One departure from general undersaturation in open ocean areas was observed as the ship passed the Humboldt Current (30° to 35° S), a feature that was included into the estimate for coastal waters. Were this feature considered as part of the open ocean, the global net saturation anomaly would become -13%. Similarly, if the inland passage were included as part of the coastal region, the mean net anomaly would become -17%.

Although the data presented here were collected along a cruise track selected for its coverage of major oceanographic regions during austral summer and boreal winter, they still represent only the East Pacific ocean. Hence, these averages might be limited in their ability to represent global oceanic regimes. However, until larger areas of the ocean are explored, these are the best estimates we can currently derive.

The flux of CH₃Br across the air-sea interface, corrected for physical effects (19), should equal the sum of in situ loss and production in surface waters. If we assume that the saturation anomaly of CFC-11 (CFC-11) is adequate in correcting for these effects, then, under steady-state conditions, the net flux of CH₃Br that is required to maintain the observed net saturation anomaly (in moles per square meter per day) is given by where F_net, P, L, K_w, and H_g are defined in Table 2, p_a is the partial pressure of the gas in the atmosphere, and and CFC-11 are saturation anomalies (in percent). Consequently, aquatic degradation exceeds aquatic production where the net saturation anomaly is negative and vice versa. The sum of the degradation rates for hydrolysis and aquatic displacement of Br in CH₃Br by Cl- ranges from less than 1% per day in 5°C seawater to greater than 30% per day in 30°C seawater and can be calculated with reasonable certainty from water temperature alone (22). Production of CH₃Br (P), then, is estimated as the difference between net flux (F_net) and loss (L). This balance of production and loss in the ocean determines the saturation anomaly, hence the net emission, of CH₃Br. From data presented here, we calculate a global net flux of -12.6 ą 6.3 Gg year-1 (Table 2). Table 2. Calculated coefficients and fluxes for CH₃Br in three main marine regions. The weighting factor is the fraction of global ocean coverage. The open ocean was estimated to occupy 80% of all oceans, while coastal waters and upwelling regions were assumed to represent each 10% (21) of the global ocean area of 361x10¹² m² (24). All results are means derived from single data points as observed during the cruise or are calculated from those. Hg, average reciprocal solubility for CH₃Br; Kw, average air-sea exchange coefficient for CH₃Br after Wanninkhof (28); Fnet; net ocean-air flux (the conversion factor into Gigamoles per year is 1/95) after Eq. 1; L, absolute loss (negative), the sum of nucleophilic displacement by Cl^- ions, neutral hydrolysis, and downward removal (1, 10, 22); P, absolute in situ production, the difference of net flux and in situ consumption. The global net saturation anomaly was derived from weighting the numbers in row 1 to 3 with the weighting factor. [If cell contents wraps into next line or doesn't display correctly, adjust your browser to a wider window!] Region Weighting factor dCH₃Br - dCFC-11 (%) H_g (m³ atm mol^-1) K_w (m d^-1) F_net (Gg y^-1) L (Gg y^-1) P global (Gg y^-1) P per unit area (g m^-2 y^-1) Open ocean {1, 5} 0.8 -24.3% 6.67 x10^-3 4.9 -16.15 -121.8 +105.6 3.7 x10^-4 Coastal {2, 6} 0.1 37.1% 5.64 x10^-3 4.2 3.47 -14.5 +18.0 5.0 x10^-4 Upwelling {3, 4} 0.1 0.65% 7.69 x10^-3 3.9 0.04 -27.6 +27.6 7.6 x10^-4 Global -15.7% -12.64 -163.9 +151.2

An observed natural variability of only 6.3% or ~0.6 ppt in atmospheric data (23) indicates that the interhemispheric difference of 2.6 ppt is significant at a >99% confidence level and that there is no observable influence of oceanic concentrations on the mixing ratios in the lower marine boundary layer. Also, the interhemispheric ratio is indicative of significantly greater NH emissions or SH sinks for atmospheric CH₃Br. With equally distributed global sinks, this ratio would imply that NH emissions are greater by a factor of 2.4 than SH emissions. Most of the coastal areas, where CH₃Br apparently is emitted to the atmosphere, are located in the NH. However, the SH oceans cover 1.3 times the area of NH oceans (24). Hence, an uneven distribution of oceanic sources and sinks may cause part of the observed atmospheric gradient. With 80% of all coastal waters located in the NH, we estimate the net flux into the SH oceans to be 1.7 times larger than that for NH oceans. The combined effects could account for about 0.2 ppt, or 8%, of the observed interhemispheric difference, which still suggests that two-thirds of the emissions external to the oceans must emanate from the Northern Hemisphere.

Because a net saturation anomaly results from both production and loss terms in the surface water (Eq 1. and Table 2.), the loss term must be separated from production to compute the partial lifetime of atmospheric CH₃Br with respect to the ocean (1). From published degradation rates for CH₃Br (22), along with measurements of sea surface temperature and wind speed taken during the cruise, this approach yields a partial lifetime of 2.9 to 3.6 years for the three regions with a best global estimate of 3.0 years (25). Combined with a lifetime of 1.7 to 2.1 years for the atmospheric removal by OH (12, 13) and stratospheric processes (26), this yields a total lifetime of 1.16 years (1.07 to 1.33 years ) for CH₃Br in the atmosphere. This lifetime estimate, along with assumptions for calculating the ozone depletion potential (ODP) as described in Mellouki et al. (12), corresponds to an ODP for CH₃Br of 0.44, which is considerably smaller than the estimate of 0.65 based on atmospheric removal of CH₃Br alone.

From the mean mole fraction of 9.8 ppt and a lifetime of 1.7 to 2.1 years due to atmospheric processes, we calculate an annual sink of 78 to 97 Gg. Adding oceanic degradation of 46 to 57 Gg y^-1, computed from the 2.9 to 3.6 year oceanic lifetime, yields a total sink of 124 to 154 Gg year^-1, with a best estimate, based upon the combined lifetime of 1.16 y, of 142 Gg year-1. These losses, plus 0 to 5 Gg year-1 for growth (3, 11), must be balanced by all sources, including that from the oceans. Conversely, adding the observed, air-to-sea flux of 6 to 19 Gg y^-1 into the ocean to tropospheric and stratospheric losses, leaves 84 to 116 Gg y^-1 for emissions external to the oceans, such as those due to agriculture, industrial use, and biomass burning.

Note added in proof: An expedition across the Atlantic ocean between 53° N and 47° S at ~30° W, conducted during the review of this report, yielded results that are both qualitatively and quantitatively consistent with findings presented here (27). REFERENCES AND NOTES 1 J.H. Butler, Geophys. Res. Lett. 21, 185 (1994). 2 H. Singh and M. Kanakidou, ibid, 20, 133 (1993). 3 M.A.K. Khalil, R.A. Rasmussen, R. Gunawardena, J. Geophys. Res. 98D, 2887 (1993). 4 C.E. Reeves and S.A. Penkett, Geophys. Res.Lett. 20, 1563 (1993). 5 D.L. Albritton and R.T Watson, in: Methyl Bromide: Its Atmospheric Science Technology and Economics, Montreal Protocol Assessment Supplement, ed. R.T. Watson, D.L. Albritton, S.O. Anderson, S. Lee-Bapty, United Nations Environmental Programme (UNEP), Nairobi, Kenia (1992), p. 3-18. 6 Scientific Assessment of Ozone Depletion: 1991, (World Meteorological Organization, Geneva 1992), p. 4.15. 7 S.M. Schauffler et al., Geophys. Res. Lett. 20, 2567 (1993). 8 K. Yagi, J. Williams, N.-Y. Wang, R.J. Cicerone, Proc. Natl. Acad. Sci. U.S.A. 90, 8420 (1993). 9 S. Manö and M.O. Andreae, Science 263, 1255 (1994). 10 H.B. Singh, L.J. Salas, R.E. Stiles, J. Geophys. Res. 88D, 3684 (1983). 11 R.J. Cicerone, L.E. Heidt, W.H. Pollock, ibid. 93D, 3745 (1988). 12 A. Mellouki et al., Geophys. Res.Lett. 19, 2059 (1992). 13 Z. Zhang et al. Geophys. Res. Lett. 19, 2413 (1992). 14 Designed by R.F. Weiss of Scripps Institute of Oceanography, see J.H. Butler et al., NOAA Data Rep. ERL ARL-16, (NOAA, Boulder, CO 1988). 15 J.H. Butler, J.W. Elkins, T.M. Thompson, B.D. Hall, J. Geophys. Res. 96D, 22347 (1991). 16 For shipboard measurements, ~200 ml of sample air was preconcentrated onto a 10 cm by 0.53 mm Al₂O₃/KCl coated fused silica trap at -50°C after passing a P₂O₅-coated dryer (Sicapent). By heating the trap to 110°C, the sample was injected onto a 30m by 0.25mm DB-5 column, which was temperature programmed from 20° to 180°C. The column was further heated to 220°C for another 15 min to purge out highly retained species. Trapping efficiency, dryer reliability and linearity were verified on another, similar instrument. 17 The hemispheric mole fraction was determined from Here, is the hemispheric mole fraction, are the measured atmospheric mole fractions of the compound, and are the total number of moles in the atmosphere and hemisphere (na =1.77x10²⁰ moles), and i and i-1 are the latitudes at the beginning and the end of each incremental interval. Constant mixing ratios were assumed between 45.3° and the north pole (11.55 ppt) and between -53.3° and the south pole (8.25 ppt). 18 Changing the ITCZ from 0° to 5°N, the hemispheric means change from 11.1 to 11.3 ppt for the NH, and from 8.5 to 8.6 ppt for the SH. 19 The saturation anomaly is defined as the percent departure of the observed dissolved amount from equilibrium: where p_w and p_a are the partial pressures of the gas in water and air. Because inert compounds like CFC-11 do not react with seawater, the departures of their surface partial pressures from equilibrium are directly attributable to physical processes such as warming and cooling, advection, or mixing, and can be used under certain conditions to factor out physical effects for gases that are not conservative in the surface ocean (14). In Eq. 1, a net saturation anomaly is calculated by subtracting the CFC-11 saturation anomaly from the observed anomaly of CH₃Br. 20 H.J. McLellan, Elements of Physical Oceanography (Pergamon Press Inc., Oxford 1965). Nitrous oxide is a strong indicator for deep water upwelling regions and was measured during this cruise in both air and surface water with a GC/ECD. 21 J.H. Ryther, Science 166, 72 (1969) assumes 9.9% for coastal waters including offshore areas of high productivity and 0.1% for regions of intense upwelling. H.W. Menard and S.M. Smith, J. Geophys. Res. 71, 4305 (1966) show that 7.5% of the ocean is shallower than 200 m, the classical definition of coastal waters. Because we found positive CH₃Br saturation anomalies in waters deeper than 200m and have included less intense areas of upwelling in our estimate, we assigned 10% of the ocean to each of these regions. 22 S. Elliot and F.S. Rowland, Geophys. Res. Lett. 20, 1043 (1993). W. Mabey and T. Mill, J. Phys. Chem. Ref. Data 7, 383 (1978), and reference 10. 23 One residual standard deviation from a fit of atmospheric data to a locally weighted, statistical smoothing (LOESS) algorithm was 0.6 ppt (W.S. Cleveland, J. Amer. Stat. Assoc. 74, 829 (1979)). 24 E. Kossina,Inst. f. Meereskunde Veröff Geogr. Naturwiss. 9 (1921). 25 The partial atmospheric lifetime with respect to oceanic loss, tau, is computed as with the area of the ocean A=361x10¹² m², the mean depth of the mixed surface water layer z=75 m (see Y.-H. Li, T.-H. Peng, W.S. Broecker, H.G. Östlund, Tellus 36B, 212 (1984)), the reciprocal solubility H for each oceanic regime (Table 2), the mass of the atmosphere M=1.77x10²⁰ moles, the air-sea exchange coefficient K_w (Table 2), and the in situ degradation rate k_s (1). The term for downward transport is calculated from an in-situ degredation rate k_z at an estimated thermocline temperature and the diffusivity D_z given by Li et al. . 26 Global Ozone Research and Monitoring Project (WMO Report 25, World Meteorological Organization, Geneva 1991). Mellouki et al. (12) included stratospheric losses in their estimate, while Zhang et al. (13) did not. We used a stratospheric lifetime of 50 years from the WMO report to adjust the number of Zhang et al. 27 J.M. Lobert et al. unpublished results from Polarstern cruise ANT XII/1, October to November 1994. 28 R. Wanninkhof, J. Geophys. Res. 97C, 7373 (1992). 29 This study was funded by the Methyl Bromide Global Coalition and the Atmospheric Chemistry Project of NOAA's Climate and Global Change Research Program. We thank M.R. Nowick for his electrical expertise, E. Saltzman for critical review, the NOAA/PMEL group for their support, and the helpful crew of the NOAA Ship Discoverer. July 1994; accepted 27 December 1994

Region	SST (°C)	V (m/s)	dCH₃Br (%)	dCFC11 (%)
Open ocean {1, 5}	21.9	7.1	-21.8	2.5
Coastal {2, 6}	17.6	7.2	40.3	2.9
Upwelling {3, 4}	26.1	6.3	1.4	0.6

Region	Weighting factor	dCH₃Br - dCFC-11 (%)	H_g (m³ atm mol^-1)	K_w (m d^-1)	F_net (Gg y^-1)	L (Gg y^-1)	P global (Gg y^-1)	P per unit area (g m^-2 y^-1)
Open ocean {1, 5}	0.8	-24.3%	6.67 x10^-3	4.9	-16.15	-121.8	+105.6	3.7 x10^-4
Coastal {2, 6}	0.1	37.1%	5.64 x10^-3	4.2	3.47	-14.5	+18.0	5.0 x10^-4
Upwelling {3, 4}	0.1	0.65%	7.69 x10^-3	3.9	0.04	-27.6	+27.6	7.6 x10^-4
Global		-15.7%			-12.64	-163.9	+151.2