http://en.wikipedia.org/wiki/Atmosphere_of_Jupiter wrote:
<<Circulation in Jupiter's atmosphere is markedly different from that in the atmosphere of Earth. The interior of Jupiter is fluid and lacks any solid surface. Therefore, convection may occur throughout the planet's outer molecular envelope. As of 2008, a comprehensive theory of the dynamics of the Jovian atmosphere has not been developed. Any such theory needs to explain the following facts: the existence of narrow stable bands and jets that are symmetric relative to Jupiter's equator, the strong prograde (super-rotating) jet observed at the equator, the difference between zones and belts, and the origin and persistence of large vortices such as the Great Red Spot.
The theories regarding the dynamics of the Jovian atmosphere can be broadly divided into two classes: shallow and deep. The former hold that the observed circulation is largely confined to a thin outer (weather) layer of the planet, which overlays the stable interior. The latter hypothesis postulates that the observed atmospheric flows are only a surface manifestation of deeply rooted circulation in the outer molecular envelope of Jupiter. As both theories have their own successes and failures, many planetary scientists actually think that the true theory will include elements of both models.
Shallow models
The first attempts to explain Jovian atmospheric dynamics date back to the 1960s. They were partly based on terrestrial meteorology, which was well developed at that time. Those shallow models assumed that the jets on Jupiter are driven by small scale turbulence, which is in turn maintained by moist convection in the outer layer of the atmosphere (above the water clouds). The moist convection is a phenomenon related to the condensation and evaporation of water and is one of the major drivers of terrestrial weather. The production of the jets in this model is related to a well-known property of two dimensional turbulence—the so-called inverse cascade, in which small turbulent structures (vortices) merge to form larger ones. The finite size of the planet means that the cascade can not produce structures larger than some characteristic scale, which for Jupiter is called the Rhines scale. Its existence is connected to production of Rossby waves. This process works as follows: when the largest turbulent structures reach a certain size, the energy begins to flow into Rossby waves instead of larger structures, and the inverse cascade stops. Since on the spherical rapidly rotating planet the dispersion relation of the Rossby waves is anisotropic, the Rhines scale in the direction parallel to the equator is larger than in the direction orthogonal to it. The ultimate result of the process described above is production of large scale elongated structures, which are parallel to the equator. The meridional extent of them appears to match the actual width of jets. Therefore in shallow models vortices actually feed the jets and should disappear by merging into them.
While these weather–layer models can successfully explain the existence of a dozen narrow jets, they have serious problems.
A glaring failure of the model is the prograde (super-rotating) equatorial jet: with some rare exceptions shallow models produce a strong retrograde (subrotating) jet, contrary to observations. In addition, the jets tend to be unstable and can disappear over time. Shallow models cannot explain how the observed atmospheric flows on Jupiter violate stability criteria. More elaborated multilayer versions of weather–layer models produce more stable circulation, but many problems persist. Meanwhile, the Galileo probe found that the winds on Jupiter extend well below the water clouds at 5–7 bar and do not show any evidence of decay down to 22 bar pressure level, which implies that circulation in the Jovian atmosphere may in fact be deep.
Deep models
The deep model was first proposed by Busse in 1976. His model was based on another well-known feature of fluid mechanics, the Taylor-Proudman theorem. It holds that in any fast-rotating barotropic ideal liquid, the flows are organized in a series of cylinders parallel to the rotational axis. The conditions of the theorem are probably met in the fluid Jovian interior. Therefore the planet's molecular hydrogen mantle may be divided into a number of cylinders, each cylinder having a circulation independent of the others. Those latitudes where the cylinders' outer and inner boundaries intersect with the visible surface of the planet correspond to the jets; the cylinders themselves are observed as zones and belts.
The deep model easily explains the strong prograde jet observed at the equator of Jupiter; the jets it produces are stable and do not obey the 2D stability criterion.
However it has major difficulties; it produces a very small number of broad jets, and realistic simulations of 3D flows are not possible as of 2008, meaning that the simplified models used to justify deep circulation may fail to catch important aspects of the fluid dynamics within Jupiter. One model published in 2004 successfully reproduced the Jovian band-jet structure. It assumed that the molecular hydrogen mantle is thinner than in all other models; occupying only the outer 10% of Jupiter's radius. In standard models of the Jovian interior, the mantle comprises the outer 20–30%. The driving of deep circulation is another problem. In fact, the deep flows can be caused both by shallow forces (moist convection, for instance) or by deep planet-wide convection that transports heat out of the Jovian interior. Which of these mechanisms is more important is not clear yet.>>