Saturation-anomaly measurements of methyl bromide (CH3Br) were made in the eastern Pacific (3/30–4/27, 2010) and the eastern Atlantic (10/25–11/26, 2010) to assess the oceanic saturation state as the phaseout of fumigation - non-Quarantine and Pre-Shipment (non-QPS) uses of CH3Br nears completion and atmospheric concentrations continue to decline. These cruises occurred 16 years after the Bromine Latitudinal Air-Sea Transect (BLAST) cruises, which were conducted in the same regions and first established a global oceanic net sink of −12.6 Gg yr−1 for atmospheric CH3Br in 1994. Results from this study suggest saturation anomalies of CH3Br in the surface ocean have become less negative than those observed 16 years ago as the atmospheric burden has declined over the past decade. The global net sea-to-air flux was estimated at 0 to 3 Gg yr−1 in 2010, suggesting that the ocean may become a net small source to atmospheric CH3Br. There are no significant differences between this study and previous studies for measured biological loss rate constants and calculated annual production rates, suggesting that annual production rates and biological degradation rate constants for CH3Br in the surface ocean have likely remained relatively constant over the past 16 years. When including the biological loss rate constants from this study and all previous studies, the mean global biological loss rate constant is constrained to 0.05 ± 0.01 d−1 (at a 95% confidence level). Combining chemical and eddy degradation rate constants, and using an updated gas transfer velocity, we estimate the CH3Br partial atmospheric lifetime with respect to oceanic loss to be 3.1 (2.3 to 5.0) years. Although the new partial atmospheric lifetime is about 1.3 years longer than the best prior estimate, it does not change the overall atmospheric lifetime of CH3Br, 0.8 (0.7–0.9) years.
 Methyl bromide (CH3Br), an important ozone-depleting substance (ODS), contributes about 34% of the total stratospheric bromine [Montzka et al., 2011]. As the abundances and emissions of most ODSs decrease, the total tropospheric burden of organic chlorine and organic bromine are declining. Unlike other ODSs controlled by the Montreal Protocol, CH3Br has both anthropogenic and natural sources. Since a significant portion of CH3Br is emitted from natural sources [Montzka et al., 2011; Yvon-Lewis et al., 2009], the relative importance of natural CH3Br in stratospheric ozone depletion will increase with the declining anthropogenic chlorine and bromine sources.
 Because of the implementation of the Montreal Protocol and its amendments which called for the phaseout of fumigation - non-Quarantine and Pre-Shipment (non-QPS) uses of CH3Br, the atmospheric mixing ratio of CH3Br has been declining [Montzka et al., 2011]. This should lead to an increase in the saturation state of CH3Br in the surface ocean, assuming surface ocean annual production rates, and biological, chemical and eddy degradation rate constants remain the same as they were before the phaseout [Butler, 1994; Yvon-Lewis et al., 2009]. Since chemical and eddy degradation rate constants are a function of salinity, sea surface temperature (SST), thermocline temperature and mixed layer depth, it is easy to assume that they have not changed significantly since 1994. Annual production rates and biological degradation rate constants are unlikely to have changed since 1994, but this is less certain.
 In this study, we selected the cruise tracks similar to those covered during the Bromine Latitudinal Air-Sea Transect I and II (BLAST I and II) cruises (Figure 1) to determine the current global saturation state of CH3Br in the surface ocean near the end of its fumigation-non-QPS phaseout. Another goal of this study was to assess the validity of the assumption in the prior modeling studies [Butler, 1994; Yvon-Lewis et al., 2009], that annual production rates and biological degradation rate constants remain constant over time.
 The Halocarbon Air-Sea Transect - Pacific/Atlantic (HalocAST - P/A) cruises were conducted aboardR/VThomas G. Thompson and the FS Polarstern, respectively. The HalocAST - P cruise, which had an almost identical cruise track to BLAST I (1/26–2/28, 1994), started from Punta Arenas, Chile, on 30 March, 2010 and ended in Seattle, Washington, U.S., on 27 April, 2010 (Figure 1). The HalocAST-A cruise, which covered a similar latitudinal range as BLAST II (10/18–11/21, 1994) but in the eastern Atlantic, departed from Bremerhaven, Germany, on 25 October, 2010 and arrived in Cape Town, South Africa, on 26 November, 2010 (Figure 1).
 Continuous underway salinity, sea surface temperature (SST), wind speed, wind direction, air temperature and relative humidity data were collected for both cruises. Additional measurements include halocarbon air-sea measurements (including CH3Br and other 19 halocarbon compounds), CH3Br degradation rate constant measurements and plant pigment measurements.
 CH3Br and a suite of other halocarbons were measured continuously in air and surface seawater (4 m below sea level) using a gas chromatograph with a mass spectrometer (GC-MS), equipped with a Weiss-type equilibrator. The details of the analytical system are described inHu et al. . The only difference from the old analytical system [Hu et al., 2010] is that the GC column was changed from a DB-VRX (i.d. 0.25 mm; length 60 m; film 1.4 μm) to a narrow bore DB-VRX (i.d. 0.18 mm; length 40 m; film 1.0 μm) column prior to the HalocAST - P/A cruises. The new column allowed for better separation and shorter chromatograms. The instrument was calibrated using two whole-air standards which were calibrated against a whole-air standard from NOAA/ESRL Global Monitoring Division that was calibrated with the NOAA-03 scale (http://www.esrl.noaa.gov/gmd/ccl/scales.html). The reported concentrations in the air or surface seawater are expressed as dry air mole fractions (parts-per-trillion, ppt) and equilibrated dry air mole fractions (ppt). The instrumental precision for CH3Br was 4.7% (1σ) during HalocAST-P and 0.8% (1σ) during HalocAST-A. A better precision during HalocAST-A was achieved by switching to a different mass spectrometer, as the one used during HalocAST-P had a persistent filament problem that increased noise.
 Measurements of CH3Br degradation rates were conducted in the eastern Atlantic during HalocAST-A. Samples were collected daily between 1200 and 1300 (local time) using the flow-through system on the ship. The inlet was located 4 m below the sea surface. The flow-through system was flushed with seawater continuously. Each sample was divided into 2 to 4 aliquots. One was filtered through 0.2 μm MediaKap Hollow Fiber Media Filter and used to determine the chemical degradation rate. Another one to three aliquots were passed through a 63 μm pore size mesh to remove large particles and were then used to measure the total degradation rate. The biological degradation rate was determined by the difference between the total degradation and chemical rates. Degradation rate constants were measured by a stable isotope incubation technique described by King and Saltzman , Tokarczyk and Saltzman , and Tokarczyk et al. . An isotopic fractionation factor of 12k/13k = 1.074 [King and Saltzman, 1997] was used to correct the measured 13C rate constant to 12C rate constant. The uncertainty of the measurement was <0.01 d−1 and the precision between aliquots was 0.01 to 0.06 (mean: 0.03) d−1.
 All plant pigment samples were filtered through GF/F filters (nominal pore size = 0.7 μm), stored in a −80°C freezer on-board ship and brought back to the laboratory for analysis. Pigments were extracted according to the methods ofWright et al. . The extracted pigments were analyzed using a Waters HPLC (high-performance liquid chromatograph) with a 996 Photodiode array detector and a Shimadzu RF 535 Fluorescence detector (excitation set at 440 nm and emission set at 660 nm). The pigments were separated on a reversed-phase Alltech Adsorbosphere C18 column (5 μm, 250 mm × 4.6 mm i.d.) using the gradient flow described by Chen et al. . A total of 18 dominant pigments, including total chlorophyll a (chlorophyll a + divinylchlorophyll a), chlorophyll b, c2 and c3, total carotene (α + β), peridinin, 19-butanoyloxyfucoxanthin, fucoxanthin, 19-hexanoyloxyfucoxanthin, prasinoxanthin, pheophorbidea, violaxanthin, diadinoxanthin, alloxanthin, diatoxanthin, lutein, pheophytin a, and zeaxanthin, was measured with a detection limit ≤1.0 nmol L−1 and an average precision of 4.0% (1σ).
3. Results and Discussion
3.1. Air and Water Concentrations and Saturation Anomalies
 Atmospheric mixing ratios of CH3Br ranged from 5.36 to 11.2 ppt (mean 7.49; sd. 0.85) for HalocAST-P (Figure 2a) and 6.28 to 9.04 ppt (mean 7.38; sd. 0.60) for HalocAST-A (Figure 3a). It was evident that the CH3Br mixing ratios in the air have decreased more in the Northern Hemisphere (NH) than those in the Southern Hemisphere (SH), when compared to the pre-phaseout (BLAST I and II) values (Figures 2a and 3a and Table 1). The inter-hemispheric ratio of CH3Br (NH/SH) was also lower than it was during the pre-phaseout BLAST cruises (Table 1). These are consistent with the atmospheric CH3Br observations from NOAA/ESRL Global Monitoring Division (see ftp://ftp.cmdl.noaa.gov/hats/methylhalides/ch3br/) and the decreasing anthropogenic emission. A larger decline of CH3Br mixing ratios in the NH than that in the SH is due to a larger reduction in the anthropogenic emissions in the NH compared to that in the SH (http://ozone.unep.org/Data_Reporting/Data_Access/).
Table 1. Hemispheric and Global Mean Atmospheric Mixing Ratios of CH3Br and the Interhemispheric Ratios (IHR) During BLAST I, BLAST II, HalocAST-P, and HalocAST-A
 Saturation anomaly (Δ%) is defined as the percent difference between the partial pressure of a trace gas in surface seawater (pw) and air (pa):
where partial pressures in the air or surface seawater were calculated by equations (2) and (3).
where χa stands for dry air mole fractions or atmospheric mixing ratios; χw stands for equilibrated dry air mole fractions for surface seawater; pt stands for surface atmospheric pressure; pvp is water vapor pressure calculated by the formula given by Weiss and Price ; and RH stands for the relative humidity. To compare CH3Br surface seawater concentrations and saturation anomalies from this study to previous studies, we used the same water mass designations of open-ocean, coastal and coastally influenced region, upwelling and inland passage asLobert et al. [1995, 1996]. In most regions, the equilibrated dry air mole fractions of CH3Br in surface seawater observed during the current study were not significantly different than those observed during BLAST I and II (Figures 2b and 3b and Table 2). Due to decreased atmospheric mixing ratios and relatively comparable dry air mole fractions in surface seawater, saturation anomalies during HalocAST became more positive than those observed during BLAST (Figures 2 and 3 and Table 2). This change is more noticeable in the open ocean, a 24.7% increase in open-ocean mean saturation anomaly between BLAST I and HalocAST-P and a 21.2% increase between BLAST II and HalocAST-A (Table 1). The change of the saturation anomalies is not only controlled by the change of atmospheric mixing ratios, it is also affected by other physical processes, i.e., surface cooling or warming, mixing of water masses and injection of air bubbles [Butler et al., 1991]. The physical effect on saturation anomalies of CH3Br could be corrected by conservative tracers, cholorofluorocarbon-11 (CFC-11) [e.g.,Hu et al., 2010; Lobert et al., 1995, 1996; Yvon-Lewis et al., 2004]. Corrected saturation anomalies of CH3Br show similar magnitudes of increase in the open ocean and upwelling areas as the uncorrected saturation anomalies (Table 2). Since the absolute solubility of CH3Br is two orders of magnitude higher than the solubility of CFC-11 [De Bruyn and Saltzman, 1997; Warner and Weiss, 1985], the CFC-11 correction of the saturation anomalies cannot completely compensate for the effect of the physical processes on the CH3Br saturation anomalies. Therefore, it is worth noting that there is an error in the CFC-11 correction and this error may become more pronounced during HalocAST when saturation anomalies of CH3Br were near zero.
Table 2. Mean Equilibrated Dry Air Mole Fractions of CH3Br in Surface Seawater, Mean Saturation Anomalies (ΔCH3Br), and Mean Corrected Saturation Anomalies (ΔCH3Br-ΔCFC-11) in Open Ocean, Coastal Ocean, and Upwelling Areas During BLAST I/II and HalocAST-P/A
 During HalocAST-A, a large increase in surface seawater CH3Br was observed between 13° S and 30° S in the eastern Atlantic. The equilibrated dry air mole fraction of CH3Br in surface seawater reached 106 ppt with 1040% of supersaturation. Correlation of surface seawater CH3Br with chlorophyll c2, c3, fucoxanthin, 19-hexanoyloxyfucoxanthin, 19-butanoyloxyfucoxantin, diadinoxanthin, peridinin, pheophorbidea and pheophytin a (r > 0.58; p = 0.00; n = 25) at a 95% confidence level and no correlation with other pigments (r < 0.32; p > 0.12; n = 25) (Figure 4) suggest that elevated CH3Br was associated with two main algal groups, prymnesiophytes and dinoflagellates [Jeffrey et al., 1997]. Laboratory culture studies [Sæmundsdóttir and Matrai, 1998; Scarratt and Moore, 1996, 1998] have shown that CH3Br is produced in both coastal and open ocean areas and that ubiquitous phytoplankton taxa (e.g., Emiliania huxleyi and Phaeocystis sp.) can produce CH3Br at significant rates. The presence of signature pigments from these two species, such as 19-hexanoyloxyfucoxanthin, chlorophyllc3, fucoxanthin, 19-butanoyloxyfucoxantin, and their accessory pigments, chlorophyllc2 and diadinoxanthin [Antajan et al., 2004; Garrido and Zapata, 1998; Llewellyn and Gibb, 2000; Schoemann et al., 2005], indicated these taxa were likely contributors to the high concentrations of CH3Br. However, we could not exclude the possibility of other prymnesiophytes and dinoflagellates species, which could also have contributed to the elevated seawater CH3Br. The most abundant pigment at this location, pheophorbide a, is a chlorophyll degradation product, which can be produced by macrocrustaceans grazing on Phaeocystis c.f. puchetii [Vernet et al., 1996], conversion of ingested chlorophyll a by macrozooplankton and microzooplankton [Goericke et al., 2000; Welschmeyer and Lorenzen, 1985], Phaeocystis autolysis, and senescence of diatoms or Phaeocystis [Head et al., 1994; Bianchi and Canuel, 2011]. A high correlation between pheophorbide a and surface seawater CH3Br (r = 0.75; p = 0.00; n = 25), along with the presence of the signature pigments of Phaeocystis sp., suggests that elevated CH3Br was in part associated with Phaeocystis sp., some of which were grazed by zooplankton, or experienced some senescence and/or autolysis.
3.2. Loss Rate Constants
 CH3Br loss rate constants were measured in the eastern Atlantic during HalocAST-A. Chemical loss rate constants, determined using filtered seawater samples, were normalized to a salinity of 35 and compared with the calculated values from the rate expression ofKing and Saltzman . Ninety percent of the measured chemical loss rate constants were within ±0.03 d−1 (or ± 20%) of the calculated rate constants.
 Biological loss rate constants, determined by subtracting the chemical loss rate constants from the total loss rate constants, were in the range of 0 to 0.24 d−1 (mean 0.09 d−1; sd. 0.06 d−1; n = 24). They contributed 0 to 73% of the total loss rate. Tokarczyk and Saltzman  measured biological loss rate constants in some of the same areas as the current study during the Gas Exchange Experiment 1998 (GasEx 98) (Figure 5). Although these two studies were 12 years apart and were conducted in different months, no significant discrepancy was observed in the measured biological degradation rate constants from the overlapped areas, suggesting that temporal variability of CH3Br biological loss rate constants may be small. Geographically, CH3Br biological loss rate constants in the eastern Atlantic were similar to those in the northern Pacific (30°–60° N) and the Southern Ocean, but higher than the biological loss rate constants observed in the Caribbean Sea, the eastern and central Pacific (10°–30° N) (Figure 5). There is not enough information available at this time to explain the spatial differences. Including results from all of previous biological loss rate constant measurements (Figure 5) [Tokarczyk and Saltzman, 2001; Tokarczyk et al., 2001, 2003] and the data from this study, we estimated a global mean biological loss rate constant of 0.05 ± 0.01 d−1 (at a 95% confidence level) for the open ocean.
3.3. Extrapolated Global Net Sea-to-Air Flux and Global Annual Production Rate of CH3Br
 The net sea-to-air flux,F (nmol m−2 d−1), was determined from the following equation:
where H is Henry's Law constant of CH3Br from De Bruyn and Saltzman  (m3 atm mol−1); kw is gas transfer velocity (m d−1); pw and pa are defined above (atm). To make a direct comparison with calculated fluxes from BLAST I and II, we used the parameterization of kw from Wanninkhof  (equation (5)).
where u stands for wind speed and Sc is the Schmidt number of CH3Br.
 According to water mass designations from Lobert et al. [1995, 1996], we divided calculated fluxes into four regions: open ocean, coastal and coastally influenced, upwelling and inland passage. Estimated net sea-to-air fluxes from the open ocean, coastal and coastally influenced region and upwelling areas in 2010 were 0.05 Gg yr−1, 2.3 Gg yr−1 and 0.1 Gg yr−1 (Table 3), respectively. The global net sea-to-air flux was 2.5 Gg yr−1 in 2010, which is a 15 Gg yr−1 increase over that observed during the BLAST cruises in 1994 [Lobert et al., 1995].
Table 3. The Global CH3Br Saturation Anomaly, the Global Net Sea-to-Air Flux, and the Global Oceanic Production Rate of CH3Bra
Wind Speed (m s−1)
Flux (Gg yr−1)
Production (Gg yr−1)
The global ocean area, 361 × 1012 m2, and the area weight are from Kossina  and Lobert et al. . The calculated saturation anomaly, flux, and production rates in open ocean excludes the data from 13°–20° S during HalocAST-A.
1.3 × 102
1.6 × 102
 An assumed constant annual production rate of CH3Br in surface ocean was used in time-dependent models to predict atmospheric CH3Br concentrations [e.g., Butler, 1994; Yvon-Lewis et al., 2009]. However, whether this assumption is true is ambiguous. Data from BLAST and HalocAST studies provide observational evidence that this assumption is very likely to be valid over the past 16 years. Here, the CH3Br production rate in the surface ocean, P (Gg yr−1), was calculated with the equation given by Yvon-Lewis et al.  and Hu et al. :
where Δ and ΔCFC-11 are saturation anomalies of CH3Br and CFC-11;kchem, kbio and are chemical, biological and eddy degradation rate constants [e.g., Butler, 1994]; Dz is the thermocline diffusivity (1 ± 0.4 cm2 s−1) [Feely et al., 2002]; z is mixed layer depth (m); kz is chemical degradation rate constant in the thermocline; both kz and kchem are functions of temperature and salinity [King and Saltzman, 1997]; Cw is CH3Br mass concentration in the surface seawater; Astands for the surface area of open ocean, coastal ocean or upwelling regions; and all other variables are defined above. For HalocAST-A, we used measured biological degradation rate constants in the calculation of production rates whereas a global mean biological degradation rate constant (0.05 (±0.06, 1 sd.) d−1) was used to calculate the production rate during HalocAST-P. The global oceanic production rate for CH3Br was estimated at 1.6 (1.4–1.8) × 102 Gg yr−1 in 2010. A previous estimate of CH3Br production rate in 1994 did not include a biological loss term [Lobert et al., 1995]. If considering a biological degradation (0.05 (±0.06, 1 sd.) d−1), the annual production rate would be 1.6 (1.3–1.9) × 102 Gg yr−1 in 1994, similar to that calculated for 2010, suggesting that the annual production rate of CH3Br in surface ocean may have remained constant over the past 16 years.
3.4. Estimating Global Oceanic Emission, Global Oceanic Uptake Rate, and Global Net Sea-to-Air Flux of CH3Br Using 1° × 1° Gridded Model
 The extrapolated annual net sea-to-air flux of CH3Br from HalocAST data, 2.5 Gg yr−1, suggests that the ocean became a net small source of atmospheric CH3Br in 2010. Since the extrapolated flux could be biased by regional in situ wind speeds and/or regional saturation anomalies from two cruises, a better approach is needed to evaluate how the spatial and temporal variability of wind speeds or surface seawater properties may affect estimates of CH3Br global fluxes. The global CH3Br net sea-to-air flux before the phaseout was estimated with the global gridded climatological wind speeds and saturation anomalies derived from an empirical relationship observed between saturation anomaly and sea surface temperature (Δ% - SST) [King et al., 2002; WMO, 2003; Yvon-Lewis and Butler, 1997]. However, the increase in saturation anomalies of CH3Br observed during this study resulted in an invalidation of this relationship (Figure 6). Simply moving the curve upward by the offset between the global mean open-ocean saturation anomaly in 2010, and that before the phaseout, 15% [King et al., 2002], can well represent the saturation anomalies from HalocAST-A, but not for those from HalocAST-P (Figure 6). So, it appears that it may not be possible to build a new meaningful Δ% - SST relationship until atmospheric CH3Br reaches a new steady state.
where kocn is the global oceanic uptake rate constant (yr−1), natm is the mass of the atmosphere (mol) and patm is the atmospheric pressure at the surface (1 atm). As mentioned above, chemical and eddy degradation rate constants are functions of temperature and salinity, which are not likely to have changed significantly since 1994. The annual production rate and biological degradation rate constants of CH3Br in the surface ocean are likely to have remained constant over the past 16 years, as shown in our results (Sections 3.2 and 3.3). Therefore, the annual emission rate of CH3Br is not likely to change significantly before and during the fumigation - non-QPS phaseout since it is a function of production rate, chemical, biological and eddy degradation rate constants (equation (7)). By substituting production rate (P) in equation (7) with equation (6)and assuming CFC-11 in surface seawater is in equilibrium with the atmosphere,equation (7) can then be expressed as follows:
where we assume that pa = χa × patm. Because saturation anomaly (Δ %) of CH3Br in the surface ocean was a function of sea surface temperature before the atmospheric CH3Br phase down, we can use this empirical relationship to calculate saturation anomalies of CH3Br and its global oceanic emission during the pre-phaseout period, and then apply that emission to 2010.
 To account for spatial and monthly variation of surface ocean properties and surface wind speeds, equations (8) and (9) are applied to a 1° × 1° gridded data set, DS279 (NOAA/GFDL Global Oceanographic Data Set Atlas, downloaded from http://dss.ucar.edu/datasets/ds279.0/), which contains monthly gridded sea surface temperature, salinity, wind speed, and mixed layer depth. The monthly global oceanic emission and monthly global oceanic uptake rate are then expressed by:
where ; i and j stand for the indices of latitude and longitude of each grid cell; m stands for the index of month; χa,1996–1998,m and χa,year,m are monthly mean atmospheric mixing ratios of CH3Br during 1996–1998 (before the CH3Br phase down) and monthly CH3Br mixing ratio in the year of interest (e.g., 2010). We then use the northern hemispheric or southern hemispheric mean mixing ratios for gridded cells in the NH or the SH. The monthly hemispheric CH3Br mixing ratios are from NOAA/ESRL Global Monitoring Division, Boulder, CO (ftp://ftp.cmdl.noaa.gov/hats/methylhalides/ch3br/flasks/CH3BR_GCMS_flask.txt).
Equations (10) and (11) are an approach used in prior studies [Yvon-Lewis and Butler, 1997; Yvon-Lewis et al., 2009] to calculate the global oceanic emission and global oceanic uptake rate before and during the atmospheric CH3Br phase down. However, these earlier studies did not consider the difference between the coastal ocean and the open ocean. This may have resulted in an underestimate on the global net sea-to-air flux since the coastal ocean is more supersaturated with CH3Br [e.g., Hu et al., 2010; Sturrock et al., 2003] compared to the open ocean [e.g., Lobert et al., 1995]. Therefore, we improved equations (10) and (11)by separating open-ocean and coastal-ocean areas and using a different approach to calculate the coastal oceanic emission, because the Δ% - SST relationship was only applicable in the open ocean. Here, we define the areas with water depths less than 200 m as coastal ocean. Coastal ocean area in each grid cell was calculated based on the bathymetric data from the 1′ × 1′ global relief database, ETOPO (http://www.ngdc.noaa.gov/mgg/gdas/gd_designagrid.html). Improved global oceanic emissions and global oceanic uptake rates were calculated with equations (12) and (13).
where Ao and Acstand for the open-oceanic and coastal-oceanic areas in gridded cells; andPc represents CH3Br production rate in the coastal ocean (nmol m−3 yr−1).
 Five scenarios were run to evaluate the effect of our improvement (equations (12) and (13)) on the estimates of global oceanic emission, global oceanic uptake rate and global net sea-to-air flux. We also examined the sensitivities of the model on the production rates of CH3Br in the coastal ocean because the production rate in the coastal ocean may be highly variable from one location to another [Hu et al., 2010; Sturrock et al., 2003], depending on the biological productivity, phytoplankton groups, influence from terrestrial transport and human impact. In scenarios 1 and 2, we used the old gridded ocean model (equations (10) and (11)) [Yvon-Lewis and Butler, 1997; Yvon-Lewis et al., 2009] and investigated the effect of an updated parameterization of gas transfer velocity on the estimated oceanic budget of CH3Br. Our results suggest that the use of an updated parameterization from Sweeney et al. , compared to using an old parameterization from Wanninkhof , yields a lower global oceanic emission, a less negative oceanic uptake rate and a more positive net sea-to-air flux (Table 4). In contrast to the first two scenarios, we separated the coastal-oceanic areas from the open-oceanic areas in scenarios 3–5 (equations (12) and (13)). In scenario 3, we assigned a uniform production rate of 0.61 nmol m−3 d−1 (the mean production rate of CH3Br from coastal areas of HalocAST and GOMECC) [Hu et al., 2010], and a biological degradation rate constant of 0.09 d−1 (the mean biological degradation rate constant observed off the coast of Florida) [King and Saltzman, 1997] to all coastal-oceanic areas, and a global mean biological degradation rate constant of 0.05 d−1in open-oceanic areas. This yields a global net sea-to-air flux of −7 Gg yr−1 during 1996 to1998 and 2 Gg yr−1 for 2010, which are 3 Gg yr−1higher than the estimated fluxes from scenarios 1 and 2 for both the pre-phaseout and the end of phaseout time periods (Table 4), suggesting it is important to consider the difference between coastal ocean and open ocean regions. Although the spatial distribution of biological degradation rate constants and production rates may affect our estimate on the oceanic emission, oceanic uptake rates and net sea-to-air fluxes, it is difficult to find a spatial pattern or seasonal variation in the surface ocean for these two parameters (Figure 5 and Hu et al. ), which can be applied to our model. Scenarios 4 and 5 test the sensitivity of our model to variation in the production rate of CH3Br in the coastal ocean. The mean production rate of CH3Br in the coastal areas of GOMECC, 0.76 nmol m−3 d−1 [Hu et al., 2010], is about three times that from the coastal areas of HalocAST, 0.25 nmol m−3 d−1. Since both studies were conducted in different regions and seasons, it is difficult to argue if one is more representative than the other. Therefore, we ran the model (equations (12) and (13)) using a coastal production rate of 0.76 nmol m−3 d−1 and 0.25 nmol m−3 d−1in scenarios 4 and 5, resulting in a global net sea-to-air flux of 3 and 0 Gg yr−1, respectively. Considering this as one of the uncertainties for the model, along with the uncertainties from the mean biological degradation rate constants (±0.01 d−1 in open ocean and ±0.02 d−1 in coastal ocean), gas transfer velocity (±32%) [Sweeney et al., 2007], the solubility (±4%) [De Bruyn and Saltzman, 1997], the mixed layer depth (±30%) [Yvon-Lewis and Butler, 2002], the global net sea-to-air flux of CH3Br in 2010 was estimated at 2 (0–3) Gg yr−1.
Table 4. Estimated Global Oceanic Emissions, Global Oceanic Uptake Rates, and Global Net Sea-to-Air Fluxes of CH3Br Before the Atmospheric CH3Br Phasedown (1996–1998) and at the End of its Phasedown (2010), with the 1° × 1° Gridded Model Described in Section 3.4a
Oceanic Emissions for 1996–1998 or 2010 (Gg yr−1)
Oceanic Uptake Rates (Gg yr−1)
Net Sea-to-Air Fluxes (Gg yr−1)
Oceanic Uptake Rates (Gg yr−1)
Net Sea-to-Air Fluxes (Gg yr−1)
Scenarios 1 and 2 use the old gridded ocean model (equations (10) and (11)) [Yvon-Lewis and Butler, 1997] with higher spatial resolution. Scenario 1 uses the parameterization of gas transfer velocity from Wanninkhof , whereas scenario 2 uses an updated parameterization from Sweeney et al. . Scenarios 3–5 use the improved gridded ocean model (equations (12) and (13)) with different production rates of CH3Br in the coastal ocean, which are 0.61 nmol m−3 d−1, 0.76 nmol m−3 d−1, and 0.25 nmol m−3 d−1, respectively.
3.5. An Improved Estimate of the Oceanic Lifetime of Atmospheric CH3Br
 The current best estimate of the partial atmospheric lifetime of CH3Br, with respect to the oceanic loss, 1.8 to 1.9 (a full range: 1.1–3.9) years [Yvon-Lewis and Butler, 1997], was based on a 2° × 2° grid of physical properties in and over the global ocean. Biological loss rate constants used in their study were based on the measurements conducted on samples only from the coast of Florida [King and Saltzman, 1997]. Here, we revised the partial atmospheric lifetime using the numerical model described in Yvon-Lewis and Butler  with modifications to address the coastal and open ocean areas separately and with a better understanding of the biological loss rate constant.
 The partial atmospheric lifetime (τocn, years) was calculated as the reciprocal of the global oceanic uptake rate constant (kocn). It is expressed as follows:
where all the variables are defined above. The best estimate of the partial atmospheric lifetime of CH3Br with respect to oceanic uptake is now estimated at 3.1 (2.3–5.0) years, which is about 1.3 years longer than the prior best estimate.
 The total atmospheric lifetime, τ, was determined from the sum of the reciprocal of each loss process:
 Saturation anomalies of CH3Br observed during this 2010 study in the eastern Pacific and the eastern Atlantic, near the end of the phase-out of fumigation-non-QPS uses of CH3Br, were less negative than those observed 16 years prior in similar regions. The global mean saturation anomalies of CH3Br in the open ocean, coastal ocean and upwelling region were positive with values of 0.3%, 29.5%, and 2.6% in 2010.
 Measured CH3Br biological loss rate constants in the eastern Atlantic ranged from 0 to 0.24 d−1, with little difference in results from the same regions examined in previous studies. When considering all previous biological loss rate constant measurements and those from this study, the mean biological loss rate constant for the open ocean is 0.05 (±0.01) d−1. Using the calculated chemical and eddy loss rate constants and the modified global mean biological loss rate constant, the estimated partial atmospheric lifetime of CH3Br is 3.1 (2.3–5.0) years, yielding an overall atmospheric lifetime for CH3Br of 0.8 (0.7–0.9) years.
 The global net sea-to-air flux ranged from 0 to 3 Gg yr−1 in 2010 based on both simple global extrapolation and a 1° × 1° grid model. Given the uncertainties, this suggests that CH3Br in the surface ocean has reached, on average, a near-equilibrium with CH3Br in the atmosphere, owing to the declining in atmospheric burden following anthropogenic emission reductions. If anthropogenic CH3Br emissions continue to decline, the atmospheric CH3Br mixing ratio will continue to decrease and the CH3Br saturation anomaly in the surface ocean should become more positive. This would result in a positive net sea-to-air flux for CH3Br.
 Calculated annual production rates of CH3Br in the surface ocean are comparable between 1994 and 2010, suggesting that the annual production rate of CH3Br in the surface ocean may have remained constant over the past 16 years. Since oceanic production rates and biological, chemical and eddy loss rate constants are relatively constant, the oceanic emission rate and the oceanic uptake rate constant will remain the same. Therefore, for those compounds with oceanic sources and sinks and changing atmospheric abundances, it is better to link their atmospheric budgets to the oceanic production rates and the uptake rate constants rather than their net fluxes.
 This project was funded by the National Science Foundation (NSF/OCE 0927874). We thank James H. Butler for his constructive comments on this study and Stephen A. Montzka for providing the atmospheric CH3Br flask data from the global monitoring stations under the NOAA/ESRL/GMD network. We also thank the captains and crews of R/V Thomas G. Thompson and FS Polarstern for their support during the cruises. We are grateful to the chief scientist on the FS Polarstern, Karl Bumke.