Superrotation and tropical waves in idealized atmospheric models

Abstract

This thesis investigates tropical variability in idealized atmospheric general circulation models. We start by exploring the parameter space of a Held-Suarez forced dynamical core to find superrotating atmospheric states. Superrotation, a state of the atmosphere where equatorial winds are westerly, is dependent on eddy momentum fluxes. We vary four fundamental model parameters and find that the thermal Rossby number (Ro) is the primary control of superrotation. At large Ro the atmosphere's tropics become dominated by Kelvin waves that flux momentum equatorward.

Next we study the meridional propagation of extratropical Rossby waves using the linearized barotropic vorticity equation. We find that the WKB approximation is good at approximating Rossby wave propagation but fails to account for significant tunneling of wave energy in evanescent regions. We show that while climatological boreal winter and spring East Pacific winds are mostly transparent to stationary waves, even minor changes in the zonal winds can produce significant reflection.

Using the insight that Rossby waves retain significant wave energy while tunneling through evanescent regions we lay out three necessary conditions for the excitation of equatorial Kelvin waves by extratropical Rossby waves: the matching of wavenumbers and frequencies and a minimizing of evanescence between the storm track and tropical waveguide. We argue that the large Ro atmosphere superrotates because wavenumber one variability is preferred and Rossby wave evanescence is minimized.

We next look at tropical variability in a seasonally varying dynamical core, where strong easterlies, rather than westerlies, dominate the tropics. We show that with sufficient seasonality the tropical easterlies are barotropically unstable and tropical eddies are generated. The eddies are a strong function of time-of-year: Kelvin waves are prevalent near equinoxes but equatorial Rossby and easterly waves dominate near solstices.

Lastly we investigate extreme global warming integrations of a full-physics atmospheric model and find that MJO-like tropical variability and superrotation become preferred at hotter temperatures, along with extreme changes in the model's cloud parameterizations. When the atmosphere strongly superrotates the convective scheme shuts off and the stratiform scheme becomes dominant, which suggests that atmospheric models may not be capable of re-creating hothouse climates.