Simulation of stratospheric tracers using an improved empirically based two‐dimensional model transport formulation

We have developed a new empirically based transport formulation for use in our Goddard Space Flight Center (GSFC) two‐dimensional chemistry and transport model. In this formulation, we consider much of the information about atmospheric transport processes available from existing data sets. This includes zonal mean temperature, zonal wind, net heating rates, and Eliassen‐Palm flux diagnostics for planetary and synoptic‐scale waves. We also account for the effects of gravity waves and equatorial Kelvin waves by utilizing previously developed parameterizations in which the zonal mean flow is constrained to observations. This scheme utilizes significantly more information compared to our previous formulation and results in simulations that are in substantially better agreement with observations. The new model transport captures much of the qualitative structure and seasonal variability observed in stratospheric long lived tracers, such as isolation of the tropics and the southern hemisphere winter polar vortex, the well‐mixed surf‐zone region of the winter subtropics and midlatitudes, and the latitudinal and seasonal variations of total ozone. Model simulations of carbon 14 and strontium 90 are in good agreement with observations, capturing the peak in mixing ratio at 20–25 km and the decrease with altitude in mixing ratio above 25 km. We also find mostly good agreement between modeled and observed age of air determined from SF 6 outside of the northern hemisphere polar vortex. However, inside the vortex, the model simulates significantly younger air compared to observations. This is consistent with the model deficiencies in simulating CH 4 in this region and illustrates the limitations of the current climatological zonal mean model formulation. The model correctly propagates the phase of the lower stratospheric seasonal cycles in 2CH 4 +H 2 O and CO 2 . The model also qualitatively captures the observed decrease in the amplitude of the stratospheric CO 2 seasonal cycle between the tropics and midlatitudes. However, the simulated seasonal amplitudes were attenuated too rapidly with altitude in the tropics. The generally good model‐measurement agreement of these tracer simulations demonstrate that a successful formulation of zonal mean transport processes can be constructed from currently available atmospheric data sets.