A. Details of MICOM and the parameterization of gas-ocean exchange. 

In MICOM, the ocean is viewed as a stack of variable-thickness isopycnic (potential density) layers, supplemented at the surface by two variable-thickness non-isopycnic layers representing the mixed layer and a fossil mixed layer. The equations are solved on a 2 degree global grid based a Mercator projection south of 60N and a bipolar projection north of 60N.  Surface wind forcing is based on ECMWF monthly climatology. Temperature and moisture fluxes are computed using the bulk formulae of Kara et al. [2000].  In order to eliminate salinity drift, a weak restoring to surface salinity [Levitus and Boyer, 1994] is applied with a relaxation timescale of 90 days. The ocean model is coupled to a rudimentary single-layer thermodynamic ice model.  The diapycnal mixing coefficient in the ocean interior is set to 0.2x10-4 m2/s. The model ocean heat-fluxes are in good agreement with observations everywhere except the Pacific, where the model slightly over-estimates the northward heat flux [Trenberth and Caron, 2001]. The model?s thermohaline circulation is maintained with time, and also compares well with observational estimates [Schmitz, 1995].  

In the model, the air-sea exchange protocol is the same as described in J.-C. Dutay et al [2002].  We have used the following thermodynamic data: 

Coefficients for computing Henry's constant for CH3CCl3 are from Gossett [1987]:

   H = exp(A-(B/(T+273))) where,
   A = 9.777; B = 4133; T is the temperature in C. 

For computing the CFC-11 piston velocity, Kw, 

   Schmidt#  = d1 + T*(d2 + (T*(d3 + (T*d4)))). where, 

   d1 = 3501.8; d2 = -210.31; d3 = 6.1851; d4 = -0.07513. 

The piston velocity, Kw is proportional to the wind-speed squared multiplied by (Schmidt#/660.)^(-0.5) and scaled by an empirically-derived constant, with a mask in sea-ice covered areas (Wanninkhof 1992, eqn 3). The air-sea gas-exchange rate, F, is then 

   F = 0.9*Kw*[C(sat)*C(surf)]

Where C(sat), the ocean saturation concentration, is equal to the atmospheric MC concentration (ppt) divided by Henry?s constant, and C(surf) is the MC concentration in the surface-ocean (mol/m(3)).  Methyl chloroform diffusivity is 0.8 that of CFC-11 [Butler et al., 1991], so the Schmidt number is assumed to be 0.9 times that of CFC-11. 

For a gas with a relatively long lifetime in the ocean (>1 yr), the air-sea flux is only modestly sensitive to most details of the air-sea parameterization (e.g. wind-speed).  The uptake is most sensitive to the solubility and to the rate of loss from the mixed layer by either mixing into the ocean interior or chemical destruction (or both).  

B.  Hydrolysis rate of methyl chloroform. 

The temperature dependent, first-order hydrolysis reaction rate of methyl chloroform, Khydrolysis, was taken from Gerkens and Franklin [1989]:

   log10(A)=13.10; E= 117.8 kJ/mole; R= 8.314*10-3 kJ mole(-1) K(-1)

   K(hydrolysis) = 10^(13.1 - 117.8 / (2.303*R*(T+273)))

No laboratory data is available below ~20C and so the calculated hydrolysis in the water masses formed at high latitude (0-3C) is a significant extrapolation.  

C.  Evaluation of the conservative properties of methyl chloroform in sea water. 

Loss of methyl chloroform in the ocean water impacts the calculations of air-sea exchange described here.  As illustrated in Figure 2, inclusion of hydrolysis loss substantially increases the net loss of methyl chloroform to the ocean and reduces the pulse from the ocean to the atmosphere following the rapid decline in the atmosphere.  Hydrolysis is a known loss process.  In principle, loss in the ocean could also occur via biological activity.  If methyl chloroform is to be used as a diagnostic of ocean transport, the loss rates for must be much better constrained than they are at the present.  

Non-biological hydrolysis:  
The hydrolysis rate of methyl chloroform is well constrained by laboratory measurements made at warm temperature but remains uncertain at the cold temperatures characteristic of the water masses formed at high latitude.  Although this does add uncertainty to the calculated magnitude of the flux of methyl chloroform out of the high latitude ocean in the late 1990s, it does not in and of itself alter our conclusion that it may be a useful tracer of ocean dynamics.  We expect that the hydrolysis rate will be much better constrained in the near future allowing substantial improvements in the calculated the loss by this mechanism in future simulations.  Several groups have developed new techniques that allow slow hydrolysis rates to be accurately measured.  

To understand the sensitivity of our calculation to the uncertainly in the hydrolysis rate at cold temperature, we have performed a test where we have changed the parameterization of the temperature dependence of this process (as calculated above) using the following kinetic parameters.  These were chosen to produce a factor of two faster or slower hydrolysis at 3C: 

   Fast Hydrolysis: log10(A)=12.60; E= 114 kJ/mole; 
   Slow Hydrolysis: log10(A)=13.65; E= 122 kJ/mole;

The results are shown in the upper panel of Fig 2. The bounds of the shaded area illustrate that the overall shape is similar but the magnitude exchange (both in and out of the ocean) is altered.  

Biological Loss:
Consumption of methyl chloroform by biological activity is much more of a concern than is the hydrolysis reaction because any biological mechanism will be difficult or impossible to simulate.  As we note in the manuscript, however, the ocean data that do exist suggest that biological loss is not fast.  Shown in Fig. S1 is a rendition of Fig. 13 of Lobert et al (5).  Unlike CCl4 which appears to have an efficient biological loss mechanism, the ratio of methyl chloroform to CFC-11 remains essentially constant throughout the hydrograph.  Clearly, more ocean measurements are required to more fully investigate biological consumption.  

D. Validation of Tracer Transport within MICOM 

Very few measurements of methyl chloroform in the ocean have been published.  To evaluate the skill of the model to represent physical uptake and subsequent subduction through the thermocline, we have had to rely primarily on measurements of CFC-11 made during WOCE.  There are, however, a few published hydrograph profiles of methyl chloroform that we describe here.  

Methyl chloroform was measured during the 1993, 1994, 1995, 1996, and 1998 WOCE repeat cruise tracks in the Labrador Sea (AR07W).  Of these, reasonable quality data were collected in 1995, 1996 and 1998 and are available on the WOCE data server (the data from 1993 and 1994 are physically unreasonable, with surface values inconsistent with the atmospheric abundance).  Fig. S2 illustrates the time series of CFC and methyl chloroform measurements and modeled tracers interpolated onto deep potential density surfaces.  For the conservative CFCs, the model is in good agreement with the highly precise and accurate measurements obtained on these cruises.  The methyl chloroform observations have a considerably larger uncertainty than the CFC-11 measurements; nonetheless, the model captures well the magnitude of the observed methyl chloroform and the downward trend with time observed in the interior Labrador Sea - a direct signature of the downturn in the atmospheric mixing ratio.

In addition to the Labrador Sea data, there are the two published hydrographs made in the Pacific during 1992 described above and one of which (station 41) is shown in Fig. S1 [Lobert et al, 1995].  They were on the P13 line, WOCE stations 41 (39N, 165E) and 54 (31N, 165E).  We have interpolated the WOCE P13 station 41 methyl chloroform data into the model coordinates (density) and plotted the comparison in Fig. S3. The model penetration of tracers is slightly too shallow in the water column compared with observations, and the model doesn?t capture all the structure in the measured methyl chloroform profile. Both stations 41 and 54 lie close to very strong N-S gradients in sea surface temperature, sea surface height, and tracers, because they fall close to the core of the Kuroshio current (Fig. S4). One reason for difficulty of capturing the observed tracer distribution in the model simulation is that this entire region is extremely eddy-rich, making the comparison of a time-averaged coarse-mesh model with snap-shot data difficult. 

The 30-40N region in the western Pacific is a difficult region for model-data comparison in general because it intersects the path of shallow subduction. One of the strengths of a pure isopycnal model is that it is possible to set the cross-isopycnal mixing to exactly zero (in a z-coordinate model there is a large amount of unavoidable numerical diffusion). Fig. S5 shows the CFC-11 distribution measured on the WOCE P13 line and simulated by MICOM. The MICOM section shows a double-lobed structure penetrating into the ocean interior: this is the signature of a recirculation gyre. Water subducted by ventilated thermocline dynamics in the high latitude (40-50N) north Pacific first gets advected to the east in the subtropical gyre, then gets advected westward at about 20N (the tracer signature is a little shallower and a little diluted here than at 40N; shallower because isopycnals are shallower at low latitudes due to Ekman pumping.) The fact that the recirculation is actually seen in the MICOM cross-section indicates that the mixing in MICOM is actually too low. If the isopycnal mixing was increased slightly, the simulations would likely be closer to the WOCE observations. As it is, this is a relatively minor point, just unlucky that the model deficiency is most clearly highlighted at 31N, one of the two N. Pacific stations at which methyl chloroform was measured.

Currently no ocean GCM can very accurately match all the WOCE tracer observations, and MICOM is just one imperfect model. However, the good agreement of the model with the observation-based global CFC inventory estimate (as discussed in the text) suggests that in a global sense, this model reasonably accurately simulates tracer uptake. 

E.  Fluxes of Methyl Chloroform. 

Calculated flux of methyl chloroform between the atmosphere and the ocean are presented in Tables S1 and S2.  The latitudinal dependence of the net flux of methyl chloroform and the inert tracer into the ocean is shown in the lower panel of Fig. 2.  Nearly all of the flux into the ocean occurs in the low- and mid-latitudes. From the mid-1990s onward, there is a net flux into the atmosphere poleward of 45S and 45N.  The contrast between methyl chloroform and the inert tracer is pronounced at both low and mid-latitudes, where much larger fluxes are calculated between 45N and 45S due to hydrolysis.  At high latitudes, in contrast, the inert tracer behaves in a very similar fashion.  


Supporting References. 

Butler, J.H., J.W. Elkins, T.M. Thompson, B.D. Hall, T.H. Swanson, and V. Koropalov (1991), Oceanic consumption of CH3CCl3: implications for tropospheric OH, J. Geophys. Res., 96, 22347-22355. 

Dutay J.-C. et al. (2002), Evaluation of ocean model ventilation with CFC-11: Comparison of 13 global ocean models, Ocean Modeling, 4, 89-120.

Gerkens R.R. and J.A. Franklin(1989)., The rate of degradation of 1,1,1-trichloroethane in water by hydrolysis and dehydrochlorination, Chemosphere, 19, 1929-1937 

Gossett J.M. (1987), Measurement of Henrys law constants for C1 and C2 chlorinated hydrocarbons, Environ. Sci. Technol, 21, 202-208. 

Kara A.B., P.A. Rochford, H.E. Hurlburt (2000), Efficient and accurate bulk parameterizations of air-sea fluxes for use in general circulation models, J. Atmos. Oceanic Tech., 17, 1421-1438. 

Levitus S., T. Boyer (1994), World Ocean Atlas, Vol. 4: Temperature: NOAA Atlas NESDIS.

Lobert J.M., T.J. Baring, J.H. Butler, S.A. Montzka, R.C. Myers, J.W. Elkins (1995), OAXTC 92: Ocean / atmosphere exchange of trace compounds 1992. Measurements of HCFC-22, CFC-11, CFC-12, CFC-113, CH3CCl3, CCl4, and N2O in the Marine air and Surface Waters of the West Pacific Ocean (03. August to 21. October 1992), NOAA Technical Memorandum ERL CMDL-9.

Schmitz W.J. (1995), On the interbasin-scale thermohaline circulation, Rev. Geophys., 33, 151-173.

Trenberth KE and J.M. Caron JM (2001), Estimates of meridional atmosphere and ocean heat transports, J. of Climate, 14, 3433-3443 

Wanninkhof (1992), Relationship between wind speed and gas exchange over the ocean, J. Geophys. Res. 97, 7373-7382