Evaluating simulations of interhemispheric transport: Interhemispheric exchange time versus SF6 age

Two recent studies using sulfur hexafluoride (SF6) observations to evaluate interhemispheric transport in two different ensembles of atmospheric chemistry models reached different conclusions on model performance. We show here that the different conclusions are due to the use of different metrics and not differences in the performance of the models. For both model ensembles, the multimodel mean interhemispheric exchange time tau(ex) agrees well with observations, but in nearly all models the SF6 age in the southern hemisphere is older than observed. This occurs because transport from the northern extratropics into the tropics is too slow in most models, and the SF6 age is more sensitive to this bias than tau(ex). Thus, simulating tau(ex)correctly does not necessarily mean that transport from northern midlatitudes into the southern hemisphere is correct. It also suggests that more attention needs to be paid to evaluating transport from northern midlatitudes into the tropics.