Tracer transport in the Martian atmosphere

Transport is crucial to understanding and reproducing the Martian dust and water cycles, and to interpreting putative methane and other trace gas (e.g. Argon) observations. However, as quantified by comparing model predictions with Argon measurements made by the Gamma Ray Spectrometer [e.g. Sprague et al., 2004, 2007], current Mars general circulation models (GCM's) appear to do a poor job at tracer transport [e.g. Nelli et al., 2007]. This invalidates a core assumption of GCM modeling in the last decade - that transport is sufficiently well treated in models that we need focus only on improving physical parameterizations, and that differences between results from different GCM's stem purely from their treatment of physical processes. If instead it is the simulated dynamical processes that need better treatment we need to move towards higher-quality numerics, e.g. based on the finite volume formulation, and introduce a more sophisticated approach to advection following work done for terrestrial chemical transport modeling. Here we present the results of non-condensable tracer transport simulations using our newly developed Mars MITgcm, which has both of the aforementioned desirable attributes: a finite volume core and access to a range of sophisticated advection schemes. Our results are encouraging in that we are able to reproduce the observed peak polar Argon enhancement factor of six (about double that attainable with most other Mars GCMs). Our diagnostics show that the time-averaged zonal-mean advection produces net increases of Argon at the winter poles, while stationary and transient eddies transport Argon away from the poles. Using less diffusive nonlinear advection schemes with flux limiters tends to produce more advective fluxes of tracers into the southern winter pole than the more diffusive linear advection schemes, resulting in a greater net increase of polar Argon abundance. We further utilize a more realistic k-distribution radiative transfer model, an improvement over the wide-band models often used in Mars GCM's, and find that the improved radiative transfer in combination with an appropriate advection scheme can quantitatively reproduce the observed Argon distribution. Our results suggest that both advection schemes and physics parameterizations are equally important in good simulation of transport-sensitive climate components, and that newer finite-volume based models (such as our MITgcm and the GFDL / NASA Ames finite volume GCM) are likely necessary for further progress in these areas.