Planetary scientists studying giant planet atmospheres continue to struggle with the most basic of questions: are the jets confined to a shallow layer within and immediately below the clouds (e.g., Hess and Panofsky, 1951), or do the cloud-level jets extend through the molecular envelopes? Or is it some hybrid of the two?
The "deep structure" scenario stems from the observation that internal heat fluxes are transported by convection, and convection homogenises entropy and density to create near-barotropic conditions in the molecular envelope. Taylor-Proudman theorem then implies that the jets will be constant on cylinders parallel to the rotation axis (Taylor columns), extending far below the cloud-decks. However, at cloud-level, the conditions are non-barotropic (temperature gradients and wind shears exist) and winds in this outer baroclinic zone would not necessarily represent the speeds of Taylor columns in the interior barotropic zone.
But even with this deep structure, there must be some depth to which strong zonal jets can no longer penetrate. Lorentz forces in the metallic hydrogen regions (i.e., the dynamo-generating region responsible for the magnetic fields) would brake strong zonal flows (Kirk and Stevenson, 1987) at pressures exceeding 1-2 Mbar. Liu et al. (2008) also suggest that the jets cannot reach the base of the molecular envelope, where a smooth transition to metallic, conducting properties would be occurring. Indeed, Liu calculated that Ohmic dissipation would be much larger than the observed luminosities of Jupiter and Saturn if the Taylor columns penetrated deeper than 96% (Jupiter) or 86% (Saturn) of the radius, and hence favoured a scenario with winds confined to the weather layer. But there are methods to reduce the Ohmic dissipation with full 3D magnetohydrodynamic calculations (e.g., Glatzmaier 2008), so the jury is still out on how deep Taylor columns could extend without Ohmic dissipation becoming a substantial problem.
Showman et al. (2006) showed that deep jets can result from both shallow forcing (e.g., jet pumping by eddy momentum flux convergence from thunderstorms or baroclinic instabilities) and deep forcing (e.g., convection in the molecular envelope). In their models, they get a baroclinic thermal wind region overlying a barotropic jet region to the base of the model, suggesting that the deep winds measured by the Galileo probe at 20 bar on Jupiter (Atkinson et al., 1998) doesn't imply a deep source of the winds. Conversely, deep forcing (e.g., waves generated by convection in the molecular envelope) could induce zonal winds at shallow depths. Del Genio et al. (2009) emphasise that deep vs. shallow structure is a different (but related) problem to the question of deep vs. shallow forcing of the zonal jets, and there are a wide variety of models out there that choose one approach or the other, each trying (often successfully) to reproduce the cloud-level wind field. But the basic question remains - how do the zonal winds vary with depth, and do they remain constant down to the expected water cloud and below?
The problem is that almost all our information (temperatures, winds, dynamical tracers) comes from the cloud-forming region, where the complex transition from barotropic interior to baroclinic outer zone is taking place. Here we believe that moist adiabatic ascent (i.e., powered by latent heat from water condensation) powers the overturning circulation and eddies power the zonal jets. But the picture remains complicated, as the typical idea of air rising in anticyclonic zones and descending in warm, dry cyclonic belts could in fact be reversed at the depth of the water cloud (zones are warm, belts are cool so that moist adiabatic ascent takes place in belts). This overturning circulation in the upper tropospheric cloud forming region, well above the deep molecular envelope, will be the topic of a future post.
Just recently, Kaspi et al. (2013) have used gravity field data from the Voyager 2 flybys of Uranus and Neptune to suggest that their atmospheric wind structure must be shallow, and they suggest that shallow forcings (e.g., moisture driven motions) are responsible for their windfields. This sort of gravitational field mapping will be performed by Juno and Cassini for Jupiter and Saturn, respectively, towards the end of this decade. Combined with microwave radiometry, this is one promising approach for peering beneath the clouds to understand how deep these zonal jets penetrate.