Denning, S. A., M. Holzer, K. R. Gurney, M. Heimann, R. M. Law, P. J. Rayner, I. Y. Fung, S-M. Fan, S. Taguchi, P. Friedlingstein, Y. Balkanski, J. Taylor, M. Maiss, and I. Levin, 1999: Three-dimensional transport and concentration of SF6: A model intercomparison study (TransCom 2). Tellus, 51B(2), 266-297.

Abstract: Sulfur hexafluoride (SF6) is an excellent tracer of large-scale atmospheric transport, because it has slowly increasing sources mostly confined to northern midlatitudes, and has a lifetime of thousands of years. We have simulated the emissions, transport, and concentration of SF6 for a 5-year period, and compared the results with atmospheric observations. In addition, we have performed an intercomparison of interhemispheric transport among 11 models to investigate the reasons for the differences among the simulations. Most of the models are reasonably successful at simulating the observed meridional gradient of SF6 in the remote marine boundary layer, though there is less agreement at continental sites. Models that compare well to observations in the remote marine boundary layer tend to systematically overestimate SF6 at continental locations in source regions, suggesting that vertical trapping rather than meridional transport may be a dominant control on the simulated meridional gradient. The vertical structure of simulated SF6 in the models supports this interpretation. Some of the models perform quite well in terms of the simulated seasonal cycle at remote locations, while others do not. Interhemispheric exchange time varies by a factor of 2 when estimated from 1-dimensional meridional profiles at the surface, as has been done for observations. The agreement among models is better when the global surface mean mole fraction is used, and better still when the full 3-dimensional mean mixing ratio is used. The ranking of the interhemispheric exchange time among the models is not sensitive to the change from station values to surface means, but is very sensitive to the change from surface means to the full 3-dimensional tracer fields. This strengthens the argument that vertical redistribution dominates over interhemispheric transport in determining the meridional gradient at the surface. Vertically integrated meridional transport in the models is divided roughly equally into transport by the mean motion, the standing eddies and the transient eddies. The vertically integrated mass flux is a good index of the degree to which resolved advection vs. parameterized diffusion accomplishes the meridional transport of SF6. Observational programs could provide a much better constraint on simulated chemical tracer transport if they included regular sampling of vertical profiles in the middle to upper troposphere. Further analysis of the SF6 simulations will focus on the subgrid-scale parameterized transports.