Baroclinic Jets on Other Jupiters and Earths

Abstract

Dynamics of baroclinic jets on extrasolar planets is studied using three-dimensional general circulation models (GCMs) which solve the traditional hydrostatic primitive equations. The focus is on: i ) baroclinic ow and instability on hot-Jupiters; ii ) detailed GCM intercomparison in a commonly used extrasolar planet setup; and, iii ) equatorial superrotation on Earth-like planets. Stability, non-linear evolution and equilibration of high-speed ageostrophic jets are studied under adiabatic condition relevant to hot-Jupiters. It is found that zonal jets can be baroclinically unstable, despite the planetary size of the Rossby deformation scale, and that high resolution is necessary to capture the process. Non-linear jet evolution is then used as a test case to assess model convergence in ve GCMs used in current hot-Jupiter simulations. The GCMs are also tested under a diabatic condition (thermal relaxation on a short timescale) similar to that used in many hot-Jupiter studies. In the latter case, in particular, the models show signi cant inter- and intra-model variability, limiting their quantitative prediction capability. Some models severely violate global angular momentum conservation. The generation of equatorial superrotation in Earth-like atmospheres, subject to \Held & Suarez-like" zonally-symmetric thermal forcing is also studied. It is shown that transition to superrotation occurs when the meridional gradient of the equilibrium surface entropy is weak in this setup. Two factors contribute to the onset of superrotation | suppression of breaking Rossby waves (generated by midlatitude baroclinic instability) that decelerate the equatorial ow, and, generation of inertial and barotropic instabilities in the equatorial region that provide the stirring to accelerate the equatorial ow. In summary, forcing condition and physical setup used in current hot-Jupiter simulations severely stretch model performance and predictive capability. Superrotation in Earth-like conditions may be common, but its strength decreases with resolution. Broadly, numerical convergence must be assessed in GCM experiments for each problem or setup considered.