Auxiliary material for Paper 2007GL032799 Use of SF5CF3 for ocean tracer release experiments David T. Ho Lamont-Doherty Earth Observatory of Columbia University, Palisades, New York, USA James R. Ledwell Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA William M. Smethie Jr. Lamont-Doherty Earth Observatory of Columbia University, Palisades, New York, USA Ho, D. T., J. R. Ledwell, and W. M. Smethie Jr. (2008), Use of SF5CF3 for ocean tracer release experiments, Geophys. Res. Lett., 35, L04602, doi:10.1029/2007GL032799. Introduction These files contain plots of atmospheric mixing ratios of CFC-11, CFC-12, and SF6, as well as their ratios, typical chromatograms from the analytical technique described for a water sample from the Santa Monica Basin, as well as a standard gas sample, detailed profiles of SF5CF3 and SF6 from the 4 sampling cruises (S1, S2, S3, and S3) to the Santa Monica Basin, as well as detailed examination of the behavior of SF5CF3 vs. SF6 during S2, S3, and S4, and more detailed model results with different plausible diapycnal diffusivities. 2007gl032799-fs01.eps. (Figure S1) Time series of a) atmospheric CFC-11 and CFC- 12 , and SF6 in the Northern Hemisphere; b) ratios of CFC-11, CFC-12, and SF6 in the Northern Hemisphere atmosphere. 2007gl032799-fs02.eps. (Figure S2) Chromatograms of SF6, SF5CF3 and CFC-12. The dotted line is 0.25 ml of a gas standard mixture (SF6 = 88.8 ppt; SF5CF3 = 127.4 ppt; CFC-12 = 115.2 ppt), and the solid line is a water sample from 775 m in the interior of the SMB taken during S3 (SF6 = 34.3 fmol/kg; SF5CF3 = 36.6 fmol/kg; CFC-12 = 908 fmol/kg). 2007gl032799-fs03.eps. (Figure S3) Map of integrating sampler tows during S1, along with the vertical profiles from each tow plotted against height above and below the injection depth. 2007gl032799-fs04a.eps. (Figure S4a). Map of sampling stations during S2, along with the corresponding vertical profiles from each station plotted against potential temperature. The green line in each vertical profile plot corresponds to the mean potential temperature of the isopycnal surface during injection in January 2005. 2007gl032799-fs04b.eps. (Figure S4b) Map of sampling stations during S2, along with the corresponding vertical profiles from each station plotted against pressure. 2007gl032799-fs05a.eps. (Figure S5a) Map of sampling stations during S3, along with the corresponding vertical profiles from each station plotted against potential temperature. The green line in each vertical profile plot corresponds to the mean potential temperature of the isopycnal surface during injection in January 2005. 2007gl032799-fs05b.eps. (Figure S5b) Map of sampling stations during S3, along with the corresponding vertical profiles from each station plotted against pressure. 2007gl032799-fs06a.eps. (Figure S6a) Map of sampling stations during S4, along with the corresponding vertical profiles from each station plotted against potential temperature. The green line in each vertical profile plot corresponds to the mean potential temperature of the isopycnal surface during injection in January 2005. 2007gl032799-fs06b.eps. (Figure S6b) Map of sampling stations during S4, along with the corresponding vertical profiles from each station plotted against pressure. 2007gl032799-fs07a.eps. (Figure S7a) Scatterplot for all samples deeper than 750 dbar from S2 of the ratio of the SF5CF3 concentration, C5, to the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg, as a function of pressure. The dashed line is the mean of these ratios, excluding samples for which C6 was less than 1 fmol/kg. There is no trend here, or if there is one, C5 is relatively low in the deep water. 2007gl032799-fs07b.eps. (Figure S7b) Scatterplot for all samples deeper than 750 dbar from S2 of the ratio of the SF5CF3 concentration, C5, to the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg, versus C6. The dashed line is the mean of these ratios, excluding samples for which C6 was less than 1 fmol/kg. The ratio appears to be relatively high for C6 greater than about 40 fmol/kg, suggesting a nonlinearity in the analysis that has not been fully accounted for. 2007gl032799-fs07c.eps. (Figure S7c) Scatterplot for all samples deeper than 750 dbar from S2 of the SF5CF3 concentration, C5, versus the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg. The straight line is a least squares fit, forced to pass through the origin. The points between 10 and 40 fmol/kg tend to lie below the line, while those beyond 40 fmol/kg lie above. 2007gl032799-fs08a.eps. (Figure S8a) Scatterplot for all samples deeper than 750 dbar from S3 of the ratio of the SF5CF3 concentration, C5, to the background corrected SFÂ6 concentration, C6 - Cb. The dashed line is the mean of these ratios, excluding samples for which C6 was less than 1 fmol/kg. There is no trend here at all. 2007gl032799-fs08b.eps. (Figure S8b) Scatterplot for all samples deeper than 750 dbar from S3 of the ratio of the SF5CF3 concentration, C5, to the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg, versus C6. The dashed line is the mean of these ratios, excluding samples for which C6 was less than 1 fmol/kg. The ratio appears to decrease with increasing C6. This trend can be made to disappear for a lower value of Cb. 2007gl032799-fs08c.eps. (Figure S8c) Scatterplot for all samples deeper than 750 dbar from S3 of the SF5CF3 concentration, C5, versus the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg. The straight line is a least squares fit, forced to pass through the origin. As in the figure to the above right, a trend is apparent with points lying above the line for low concentrations and below the line for higher concentrations. 2007gl032799-fs09a.eps. (Figure S9a) Scatterplot for all samples deeper than 750 dbar from S4 of the ratio of the SF5CF3 concentration, C5, to the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg, as a function of pressure. The dashed line is the mean of these ratios, excluding samples for which C6 was less than 1 fmol/kg. There is little trend seen, except that the deep points appear to be high. This is weak evidence for preferential settling of SF3CF3. Note, however, that such evidence is lacking for S2 and S3. 2007gl032799-fs09b.eps. (Figure S9a) Scatterplot for all samples deeper than 750 dbar from S4 of the ratio of the SF5CF3 concentration, C5, to the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg, versus C6. The dashed line is the mean of these ratios, excluding samples for which C6 was less than 1 fmol/kg. No trend is visible in this plot. 2007gl032799-fs09c.eps. (Figure S9c) Scatterplot for all samples deeper than 750 dbar from S4 of the SF5CF3 concentration, C5, versus the background corrected SFÂ6 concentration, C6 - Cb, with the background Cb, taken as 0.3 fmol/kg. The straight line is a least squares fit, forced to pass through the origin. No trend is visible in this figure. 2007gl032799-fs10a.eps. (Figure S10a). Model for mixing in the interior, S1 to S2. The red and black lines are the observed basin-wide mean concentrations for S1 and S2, respectively; solid: SF5CF3; dashed: SF6. The error bars represent the standard errors of the means based on the variance in the individual profiles. The blue lines represent fits with the model described in the paper for diffusivities of 3 x 10-5 m2/s (dotted blue lines), 4 x 10-5 m2/s (solid blue lines Đ the best fit), and 5 x 10-5 m2/s (dashed blue lines). All of the curves for S1 and S2 have been scaled by dividing by the maximum observed mean concentration of SF5CF3 for S1 and S2, respectively. 2007gl032799-fs10b.eps. (Figure S10b) Whole basin model for S2 to S3. The red and black lines are the observed basin-wide mean concentrations for S2 and S3, respectively; solid: SF5CF3; dashed: SF6. The error bars represent the standard errors of the means based on the variance in the individual profiles. The blue lines represent fits with the model described in the paper, with constant vertical mass flux below sill depth and constant diffusivity above sill depth. In the model, both the initial profile and the final profile are allowed to depart from the observed profiles, to minimize a cost function based on the error bars, some of which are shown in the figure. The diffusivity at 785 m in the three model runs shown were 0.7 x 10-4 m2/s (dotted blue lines), 1.1 x 10-4 m2/s (solid blue lines Đ considered the best fit), and 2.0 x 10-4 m2/s (dashed blue lines). All of the curves for S2 and S3 have been scaled by dividing by the maximum observed mean concentration of SF5CF3 for S2 and S3, respectively.