Planetary science

Plumbing the depths of Uranus and Neptune

An analysis of data collected by the Voyager 2 spacecraft and by ground-based telescopes limits the depths to which winds penetrate into Uranus and Neptune, informing the debate about these planets' internal structures. See Letter p.344

In an age in which space missions penetrate the far reaches of the Solar System, it seems incredible that planetary scientists still argue about the depth of atmospheric circulations inside the ice giants, Uranus and Neptune, and the gas giants, Jupiter and Saturn. But, because of the complexities of the problem, many uncertainties remain. The difficulties are compounded by the inaccessibility of the planets' interiors, which are hidden beneath dense layers of clouds, making it difficult to observe their structure and dynamics. It is therefore a great achievement for Kaspi et al.1 (page 344 of this issue) to have obtained, from careful and insightful analyses of the planets' gravitational fields, surprisingly tight constraints on how deeply the winds on Uranus and Neptune penetrate.

How do we know anything about the winds on these remote planets? The main sources of information date back to the close encounters of the Voyager 2 spacecraft with Uranus (in 1986) and Neptune (in 1989) during its grand tour of the outer Solar System. Images obtained during those encounters revealed methane and ammonia clouds whose motion could be tracked to determine wind strength at levels at which the atmospheric pressure was around 100–300 kilopascals. This showed that, despite being among the farthest flung and coldest planets in the Solar System, Uranus and Neptune actually have some of the strongest winds2, with east–west (zonal) velocities of up to 450 metres per second, compared with just 30–100 m s−1 on Earth. The source of energy for these immensely strong winds remains mysterious — it is unclear whether they are driven mainly by weak differential heating from the Sun or, at least in the case of Neptune, by heat upwelling from the deep interior.

Uranus and Neptune are composed primarily of hydrogen, helium and so-called icy materials (water, ammonia and hydrocarbons, albeit at temperatures of several thousand kelvin), so the interior of both planets is almost certainly predominantly fluid3,4. This raises the question of whether the winds observed at the 100-kPa level are confined to the outer atmospheric layers, or represent a deep-seated pattern of internal thermal convection. But how can researchers tell the difference? Direct measurements are limited to the outermost few tens of kilometres of the atmosphere, where cloud motions can be tracked. Indirect methods, such as those based on inferences and insight from numerical simulations, must therefore be adopted. However, these simulations have severe limitations because it is impossible to model conditions that even remotely approach those that prevail5 in the deep interiors of Uranus and Neptune.

On the basis of principles established by the fluid dynamicists Geoffrey Ingram Taylor and Joseph Proudman6,7, Fritz Busse noted8 that fairly weak differential motions in a rapidly rotating fluid tend to be coherent (almost invariant) along the direction parallel to the rotation axis, as long as surfaces of constant density are approximately horizontal. Inside a spherical planet, motions would therefore tend to be almost constant on cylindrical surfaces coaxial with the planetary rotation axis, if the fluid density were the same everywhere (Fig. 1a). In the presence of purely radial variations in density ρ, however, the product of ρ and the velocity U in the direction perpendicular to the rotation axis must be coherent along cylinders. So U would become much smaller inside the planet as the density increases, reducing by a factor of about 10,000 between the 100-kPa level and the deep interior of Uranus or Neptune (Fig. 1b).

Figure 1: Velocity fields in fluid planets.

In these cross-sections through a fluid planet that rotates about a vertical axis, the colour contours represent the magnitude of a hypothetical zonal-wind field, describing fluid motion for which 2Ω ·(ρU) = 0, where Ω is angular velocity of planetary rotation, ρ is fluid density, assumed to vary exponentially with radius, and U is fluid velocity in the east–west direction. Dark blue, westward flow; red-orange, eastward flow; yellow, weak-to-moderate eastward flow; pale blue and green, flow close to zero. The cross-sections represent examples of the kinds of interior flow pattern considered by Kaspi et al.1 for Uranus and Neptune. In each case, the latitudinal surface-velocity pattern is consistent with that of those planets. a, The density of this planet is essentially uniform. The velocity field therefore hardly varies along coaxial cylinders around the axis of rotation. b, The variation of density with depth for this planet is roughly the same as that of Uranus, increasing inwards by a factor of around 10,000 from the visible cloud tops. The velocity field strongly intensifies at the surface.

If latitudinal variations in density are included, the strict axial coherence of the flow can be broken by the effects of density variations along horizontal surfaces (known as a baroclinic flow). Real convective flows are likely to be at least partly baroclinic. Kaspi and colleagues have extended an earlier method9 by taking into account the density variations associated with baroclinic effects to constrain the depth to which the surface winds on Uranus and Neptune penetrate.

Variations in density that balance the variations in ρU along cylindrical surfaces will also affect the structure of a planet's gravitational field, which can be measured from small accelerations and decelerations in the trajectory of a spacecraft as it flies close to the planet. The zonal winds of Uranus and Neptune vary only gradually with latitude2, therefore the effects of dynamic density variations should manifest themselves in the low-order gravitational harmonics that represent the largest-scale variations in the planets' gravitational fields. Two of these harmonics (designated J2 and J4) have been measured, for example during the encounters of Voyager 2 with Uranus and Neptune. Kaspi et al. have determined that only a velocity pattern that is tightly confined to the outermost 1,000 km of each planet's radius can be consistent with the measured J2 and J4 harmonics. This result probably favours a shallow meteorology for both planets (although 1,000 km seems deep to us on Earth, because Earth's atmosphere spans a depth of just 100 km).

So what does this suggest for the larger cousins of these planets, Jupiter and Saturn? They also have strong zonal winds at their cloud tops, but these winds vary much more strongly with latitude than those of Uranus and Neptune. This means that any signature in the structure of their gravitational fields will manifest itself only in the high-order, rapidly varying gravitational moments (J10 and higher)10, which could not be measured during the Voyager fly-bys or in the ongoing Cassini mission.

In 2011, NASA launched its Juno mission, the main goal of which is to accurately measure the high-order gravitational moments of Jupiter from the vantage point of a low-altitude polar orbit. Kaspi and colleagues' approach will enable constraints to be placed on the depth of Jupiter's deep baroclinic flow. This might settle the long-running debate over whether Jupiter's winds are deep-seated or superficial. And when the Cassini orbiter reaches the end of its mission in 2017, plans are afoot for the spacecraft to enter a low polar orbit around Saturn (before crashing into it) to make the same measurements of the planet's gravitational field as Juno will for Jupiter. By the end of this decade, we may therefore have obtained measurements that constrain the wind-penetration depth for all four gas- and ice-giant planets.

If I were a betting man, my money would be on a shallow penetration depth for the gas giants, given the precedents of their ice-giant cousins. There are several other reasons to expect gas-giant meteorology to be confined to shallow depths11, one of which is that the fluid interiors of the gas giants become electrically conducting at depths of several thousand kilometres. This allows hydromagnetic forces to disrupt the 'Taylor–Proudman condition', breaking the associated axial coherence of the winds. However, a probe from the Galileo spacecraft found that zonal winds increase with depth near Jupiter's equator12, so the evidence is not unequivocal. I await the results from Juno and Cassini with considerable interest.


  1. 1

    Kaspi, Y., Showman, A. P., Hubbard, W. B., Aharonson, O. & Helled, R. Nature 497, 344–347 (2013).

  2. 2

    Ingersoll, A. P. Science 248, 308–315 (1990).

  3. 3

    Hubbard, W. B. & Marley, M. S. Icarus 78, 102–118 (1989).

  4. 4

    Podolak, M., Weizman, A. & Marley, M. Planet. Space Sci. 43, 1517–l522 (1995).

  5. 5

    Showman, A. P., Kaspi, Y. & Flierl, G. R. Icarus 211, 1258–1273 (2011).

  6. 6

    Taylor, G. I. Proc. R. Soc. Lond. A 93, 92–113 (1917).

  7. 7

    Proudman, J. Proc. R. Soc. Lond. A 92, 408–424 (1916).

  8. 8

    Busse, F. H. Icarus 29, 255–260 (1976).

  9. 9

    Hubbard, W. B. et al. Science 253, 648–651 (1991).

  10. 10

    Hubbard, W. B. Icarus 137, 357–359 (1999).

  11. 11

    Liu, J., Goldreich, P. M. & Stevenson, D. J. Icarus 196, 653–664 (2008).

  12. 12

    Atkinson, D. H., Ingersoll, A. P. & Seiff, A. Nature 338, 649–650 (1997).

Download references

Author information



Corresponding author

Correspondence to Peter Read.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Read, P. Plumbing the depths of Uranus and Neptune. Nature 497, 323–324 (2013).

Download citation


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.