Heating of Jupiter’s upper atmosphere above the Great Red Spot

The temperatures of giant-planet upper atmospheres at mid- to low latitudes are measured to be hundreds of degrees warmer than simulations based on solar heating alone can explain. Modelling studies that focus on additional sources of heating have been unable to resolve this major discrepancy. Equatorward transport of energy from the hot auroral regions was expected to heat the low latitudes, but models have demonstrated that auroral energy is trapped at high latitudes, a consequence of the strong Coriolis forces on rapidly rotating planets. Wave heating, driven from below, represents another potential source of upper-atmospheric heating, though initial calculations have proven inconclusive for Jupiter, largely owing to a lack of observational constraints on wave parameters. Here we report that the upper atmosphere above Jupiter’s Great Red Spot—the largest storm in the Solar System—is hundreds of degrees hotter than anywhere else on the planet. This hotspot, by process of elimination, must be heated from below, and this detection is therefore strong evidence for coupling between Jupiter’s lower and upper atmospheres, probably the result of upwardly propagating acoustic or gravity waves.

The spectrum in Fig. 1b shows strong emission features at six wavelengths, which appear prominently in the auroral regions and wane towards the equator. These are discrete ro-vibrational emission lines from H 3 + , a major ion in Jupiter's ionosphere, the charged (plasma) component of the upper atmosphere. The colour contours highlight the weaker emissions from this ion across the body of the planet. Far from a uniform intensity at low latitudes, there is a substantial intensity enhancement in all of the emission lines within the − 13° to − 27° planetocentric latitude range occupied by the GRS 9 . As seen in the coloured contours of Fig. 1b, the H 3 + emissions are isolated in wavelength, indicating that there is no continuum reflection of sunlight at the latitudes of the GRS.
The ratio between two or more emission lines can be used to derive the temperature of the emitting ions 10,11 . With the observing geometry used here, such temperatures are altitudinally averaged 'column temperatures' of H 3 + , where the majority of H 3 + at Jupiter has been observed to be located at altitudes between 600 km and 1,000 km above the 1-bar pressure level 12 . H 3 + has been demonstrated to be in quasi-local thermodynamic equilibrium throughout the majority of Jupiter's upper atmosphere, meaning that derived temperatures are representative of the co-located ionosphere and (the mostly H 2 ) thermosphere 13 . In the Methods section we detail the data reduction techniques and temperature model fitting procedures, and in Fig. 2   The difficulty in explaining the observed upper-atmospheric temperatures of the giant planets was realized more than 40 years ago 1 , and has since been termed the giant-planet "energy crisis" 2,4 . For Jupiter, only the observed temperatures within the auroral regions have been adequately explained, as the 1,000-1,400 K temperatures 14 observed there result from auroral heating mechanisms that impart 200 GW of power per hemisphere through ion-neutral collisions and Joule heating 15,16 . The low to mid-latitudes do not have such a heat source, and yet are measured to be near 800 K, which is 600 K warmer than can be accounted for by solar heating 15,17,18 . If heating does not come from above (solar heating), and cannot be produced in situ via magnetospheric interactions, then a solution is likely to be found below.
Gravity waves, generated in the lower atmosphere and breaking in the thermosphere, represent a potentially viable source of upperatmospheric heating. Previous modelling studies, however, have led to inconclusive results for Jupiter: while viscous dissipation of gravity waves in Jupiter's upper atmosphere can lead to warming of the order of 10 K, sensible heat flux divergence can also lead to cooling by a similar amount, depending on the properties of the wave 6,7 . Recent re-analysis of Galileo Probe data has shown that gravity waves impart a negligible amount of heating vertically to the stratosphere (gravity-wave motion is primarily longitudinal and latitudinal) and that heating near the thermosphere is less than 1 K per Jovian day 19 .
A more likely energy source is acoustic waves that heat from below (also via viscous dissipation); this form of heating requires vertical propagation of disturbances in the low-altitude atmosphere. Acoustic waves are produced above thunderstorms, and the subsequent waves have been modelled to heat the Jovian upper atmosphere by 10 K per day 20 and on Earth have been observed to heat the thermosphere over the Andes mountains 20,21 . On Jupiter, acoustic-wave heating has been modelled to potentially impart hundreds of degrees of heating to the upper atmosphere 22 . However, to the best of our knowledge, no such coupling between the lower and upper atmosphere has been directly observed for the outer planets, so vertical coupling has not been seriously considered as a solution to the giant-planet energy crisis.
Jupiter's GRS is the largest storm in the Solar System, spanning 22,000 km by 12,000 km in longitude and latitude, respectively. The GRS lies within the troposphere, with cloud tops reaching altitudes of 50 km, around 800 km below the H 3 + layer 9 . In Fig. 3 we show (red circles) that the pattern of H 3 + intensity seen above the GRS, when fitted to our model, gives column-averaged H 3 + temperatures of over 1,600 K, higher than anywhere else on the planet, even in the auroral region. We also fitted temperatures to a swath of longitudes away from the GRS in order to illustrate that the enhancement in temperature occurs only within this longitude band. The latitudinal variation of temperatures away from the GRS is similar to the ranges previously observed 17 , indicating that the high temperature above the GRS is localized in both latitude and longitude.
The high temperature in the northern part of the GRS provides direct observational evidence of a localized heating process. We interpret the cause of this heating to be storm-enhanced atmospheric turbulence, which arises due to the flow shear between the storm and the surrounding atmosphere. Some of these waves must then propagate vertically upwards, depositing their energy as heat through viscous dissipation. It is unknown, at present, why the two red data points at GRS latitudes (grey shaded region in Fig. 3 Fig. 2. Red circle symbols correspond to the co-addition of GRS-related spectra (that is, from the spectral image in Fig. 1b) between 239° and 253° in Jovian system III Central Meridian Longitude (CML). The GRS latitudes are indicated by the grey shading. Blue triangle symbols were derived from exposures taken in the ranges 293°-359° and 0°-82° CML, that is, longitudes well separated from the GRS, representing the 'ordinary' background conditions based on solar heating alone. The modelled temperature of the upper atmosphere for these non-auroral regions is 203 K (ref. 1). Uncertainties are standard errors of the mean. of the GRS may be much higher than derived, but only if methane is preferentially brighter in the south. However, as the H 3 + and CH 4 lines at 3.454 μ m are not separated spectrally in this work, it is not possible to conclude whether or not contamination is present.
An alternative physical explanation may relate to the relative velocities between the zonal wind and the GRS being greatest on the equatorward side of the storm: relative velocities are 75 m s −1 in the north, 15 m s −1 in the storm core, and 25 m s −1 at the poleward edge 9 . The largest relative velocities would induce the strongest flow shear, leading to the greatest turbulence and therefore the largest contribution to heating above. It is possible that evidence of such energy transfer from the lower to the upper atmosphere would be deposited en route in the intervening troposphere and upper stratosphere (0-150 km altitude), as there is a temperature enhancement of 10 K encircling the GRS at these altitudes 23,24 . However, this enhancement could also be due to the upwelling of material in the centre of the GRS, followed by increased adiabatic heating when the material downwells around the edges 24 .
The only previous map of Jovian H 3 + temperatures that contains the GRS was made using ground-based data obtained in 1993 (ref. 17). The authors of ref. 17 did not mention the GRS, as no obvious signature was present in their temperature map. However, on the basis of their temperature contours and the expected location of the GRS at the time, we estimate that there was a measured temperature enhancement of 50 K above the GRS. Such a minor temperature increase may indicate that the GRS-driven heating of Jupiter's upper atmosphere is transient, but the spatial resolution of the 1993 observations was 9,800 km per pixel (at the equator), compared with 500 km per pixel in this study. Therefore, the previous data had much cruder resolution in latitude and longitude, and any localized temperature enhancements would have been smoothed out.
In this work, the high-temperature region above the GRS is localized in latitude and longitude, indicating a large temperature gradient and perhaps a confinement by currently unknown upper-atmospheric dynamics. If wave heating driven from below is responsible for the temperatures observed in Jupiter's non-auroral upper atmosphere, then we might expect a relatively smooth temperature profile with latitude, punctuated by temperature enhancements above active storms. The GRS may then simply be the 'smoking gun' that dramatically illustrates this atmospheric coupling process, and provides the clue to solving the giant-planet energy crisis.
Online Content Methods, along with any additional Extended Data display items and Source Data, are available in the online version of the paper; references unique to these sections appear only in the online paper.