FIRST RESULTS FROM THE HUBBLE OPAL PROGRAM: JUPITER IN 2015

Table 1 summarizes the WFC3 data acquired for Jupiter on 2015 January 19. All images were acquired with the UVIS detector, whose pixel scale is about 0.04 arcsec (Dressel 2015). Filters were chosen to allow maximum cloud contrast for wind tracking (typically the red continuum wavelengths), altitude discrimination (UV and methane-gas absorption wavelengths) and color analyses (blue and green wavelengths). To generate a global map of Jupiter requires 6–7 consecutive orbits of Hubble (each 96-minute orbit allows about 52 minutes of target visibility between Earth occultations). Due to the Earth's South Atlantic Anomaly, which introduces a higher rate of high-energy particle hits on the detectors, it is sometimes necessary to skip an orbit with a more direct passage into this region of Earth's magnetic field. Hubble Space Telescope gyroscope bias reset activities also prevent blocks of more than seven consecutive orbits from being scheduled. Thus, the Jupiter data skip an orbit in-between maps. Because of overlap at the end of a seven-orbit sequence some regions were visible with an ∼10-hr separation, while a limited longitude range also had a 20-hr separation. Finally, the observations were also designed to allow for small time separations within an orbit, repeating the red filter at the beginning and end of each sequence.

All images were retrieved after the standard Hubble pipeline processing was performed. Fringing introduces photometric error due to multiple internal reflections within the detector, since silicon is increasingly transparent at long wavelengths. The photometric error is spatially variable across the image and highly dependent on the wavelength of illumination. The effect is significant only for narrowband filters at wavelengths >650 nm (Wong 2010). Standard flatfield corrections remove fringing error when the target's spectral energy distribution is the same as the continuum spectrum of the calibration lamps used to construct the flatfields themselves. However, Jupiter's spectrum is different, particularly near methane absorption bands, requiring additional correction. Our FQ889N images were corrected by about 2%, using preliminary fringing flatfields (Wong 2011) generated using the disk-averaged reflectance spectrum of Jupiter (Karkoschka 1998) convolved with the solar spectrum (Colina et al. 1996).

After corrections, each image was navigated to find the planet center using iterative ellipsoid limb fitting with Jupiter's known radii and geocentric distance and assigned proper sub-Earth and sub-solar latitude and longitude. Navigated images were then adjusted for limb darkening using a Minnaert approximation (ratio of the cosines of the incident and emission angles raised to the k power) with the k coefficients listed in Table 1, and projected onto planetographic cylindrical maps. These maps were then mosaicked in each wavelength, except F467M, which was saturated, to produce global maps for further analysis, Figures 1(a) and (b). These maps are publicly available at https://archive.stsci.edu/prepds/opal/.

Zoom In Zoom Out Reset image size

Jupiter's zonal wind profile is known to be relatively stable with time, with only small (10–20 m s⁻¹) variations at some wind jet peaks (Limaye 1989; Simon-Miller & Gierasch 2010; Asay-Davis et al. 2011). The OPAL data set will allow for the retrieval of a full two-dimensional wind field, requiring iterative automated retrievals to generate sufficient density of wind vectors. However, a quick comparison of the zonal wind field can be performed with a simple cross-correlation between the two global maps. Figure 2 shows such a correlation, computed from 360° of longitude and at every 01 of latitude, using the global maps in the F631N filter; the retrieved zonal wind is plotted at the displacement with maximum correlation (solid black line). While the profile retrieval can be further refined, the preliminary result shows very good agreement with both the Cassini (blue line, Porco et al. 2003) and Voyager (red line, Simon-Miller & Gierasch 2010) zonal wind profiles. Manual measurements (plus symbols) indicate that the changes in the zonal jets, compared with Voyager and Cassini are real.

Zoom In Zoom Out Reset image size

Notably, the peak of the 24° N jet is at about 153 m s⁻¹, between the Cassini and Voyager values of 138 and 184 m s⁻¹, respectively. Other jets' differences are due, in part, to differences in the presence of features with distinct drift rates in the different epochs. For example, near 7° S, the Hubble data show a higher zonal average, but the large South Equatorial disturbance, visible during Voyager and again during Cassini was not present in 2015; it moves more slowly that the zonal wind, and may be related to large wave phenomena (Beebe et al. 1989; Simon-Miller et al. 2012). Changes in other jets require a more detailed analysis of the cloud features, retrieval of the full 2D wind circulation, and masking of vortices and other discrete features with distinct drift rates (e.g., Asay-Davis et al. 2011; Barrado-Izagirre et al. 2013), efforts that are beyond the scope of this initial report.

As the map in Figure 1 shows, many red features are visible for spectral analysis. The Great Red Spot (GRS) still appears very orange, as it did in 2014 (Simon et al. 2014). Oval BA, to its southeast, is somewhat paler, with a whitish core and fainter orange annulus. The North Equatorial Belt (NEB) and cyclonic barges in the northern hemisphere remain the visibly reddest features on the planet (Sanchez-Lavega et al. 2013; Simon et al. 2015b). Filters in the OPAL program were specifically chosen to allow comparison of the spectral characteristics of these features with previous data sets. Figure 3 shows the 2015 OPAL data (red symbols) of the core of the GRS and reddest regions of the NEB with other data sets from 1995 to 2014. Note that the F467M filter was saturated across much of the planet, but red regions far from the saturation were usable.

Zoom In Zoom Out Reset image size

The overall UV to green absorption in the GRS remains consistent with measurements from 2014, though the 889-nm methane absorption band brightness has returned to its pre-2014 value, indicating less haze reflectivity compared with 2014. The evolution of the GRS—a color change coincident with increased haze reflectivity, followed by a decrease in haze but constant color—is distinctly different from the evolution of other Oval BA (the smaller red storm to the south and east in Figure 1), where a one time color change occurred along with no change in haze reflectivity (Wong et al. 2011). These details may prove important in determining the identity and controlling processes of red chromophores on Jupiter, as well as the evolution of large anticyclones. The NEB spectrum remains nearly constant at all wavelengths.

Hubble data in 2014 confirmed that the GRS's longitudinal width has decresed significantly since 2012 (Simon et al. 2014). In 2015 it spanned 139 of longitude and 97 of latitude. This represents a return to the normal longitude shrinkage rate of ∼019/year, while the latitude extent increased slightly over the 2014 measurements. It is currently centered near 224 S latitude, and its color indicates that there is still decreased interaction with the nearby zonal wind jets (Simon et al. 2014). The time resolution of the Hubble observations are not sufficient to determine whether the observed size/shape changes have any periodic components analogous to oscillations seen by Voyager in Neptune's Great Dark Spot (Smith et al. 1989; LeBeau & Dowling 1998).

An interesting feature in the 2015 data, however, is the internal GRS appearance. The core region, usually the most prominent at violet wavelengths, has been gradually shrinking, concurrent with its overall size changes (see Simon-Miller et al. 2002, Figure 3). By 2014, this separate core was largely absent (Simon et al. 2014), and, in 2015, filamentary structures can be seen spiralling near the core, Figure 4. Over 10 hours, the motion in this structure is apparent. It was not possible to retrieve a detailed velocity field in 2014, due to Ganymede's shadow in the region, but the velocities that were measured were consistent with previous measurements. Preliminary manual measurements confirm velocities in the range of 100–150 m s⁻¹, as expected, but the OPAL data will enable later detailed flow field mapping.

Zoom In Zoom Out Reset image size

The OPAL data also show a feature not previously observed in Hubble images, a small-scale wave in the NEB, see Figures 5 and 6. These fine-scale waves reside in the cyclonic region of the wind field, below several anticyclones to the north and at the same latitudes as a series of small cyclonic cells, as denoted in Figure 5. They are visible in both global maps, and at all continuum wavelengths from 343 to 658 nm; they may be visible at 275 and 889-nm, though at very low contrast levels. The features have a wavelength of about 1° (1200 km), larger than the 300 km mesoscale waves seen in Voyager and New Horizons data (Hunt & Muller 1979; Reuter et al. 2007; Simon et al. 2015a).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

As shown in Figure 6, these waves span ∼2°–3° of latitude covering the cyclonic flank of the westward jet that peaks at 17° N. Although the wave is visible in multiple images, it is difficult to identify the same wave crests from frame to frame in large time separations, even in the filter with best contrast, F395N, Figure 6 middle and bottom panels. Wave crest identification is not unique over a 20-hr separation and no detectable wave motions were possible. Short 1-hr time separations are too short to confirm velocities, given the large velocity uncertainties over such a short time period at the available spatial resolution.

This type of wave is similar in appearance to terrestrial wave formation in a baroclinic instability (see Holton 1992, Figure 6.5). On the Earth, baroclinic instabilities explain mid- to high-latitude cyclone generation, and waves form along with cyclonic features (Charney 1947; Eady 1949). On the giant planets, such instabilities may play a role in wind jet formation (e.g., Kaspi & Flierl 2007). Using the Earth analogy, we can examine the characteristics of the OPAL wave for consistency with a baroclinic wave.

One diagnostic is the meridional size of the wave features. Stone (1969) showed that for most baroclinic waves, the meridional scale should be on the order of the radius of deformation. However, Simmons (1974) showed that if the jet width scale is larger than the deformation radius, the wave's meridional extent can be between the two in scale.

Gierasch et al. (1979) modeled Jupiter's atmosphere using a nearly adiabatic deep layer beneath the clouds. Unlike the Eady (1949) terrestrial model with a rigid lower boundary, this scheme had drastically reduced growth rates of baroclinic instabilities due to the influence of the deep neutrally stable layer. Conrath et al. (1981) examined this again and found that for westward jets that are broad (where the deformation scale is much smaller than the jet width), the Eady modes are still suppressed, but Charney modes (Charney 1947) are allowed and the disturbance can be a mix of barotropic and baroclinic instabilities if deformation radius and jet width are comparable. The main difference between the Eady and Charney modes is the influence of the gradient of the Coriolis parameter, which makes Charney mode growth less dependent on the bottom boundary condition. Conrath et al. (1981) hypothesized that the jet widths in Jupiter's atmosphere, at least 3–10 times the tropospheric deformation radius, would enable transfer of energy from eddies to jets. Instabilities are baroclinic within this range.

The atmospheric radius of deformation is defined as

$$\begin{eqnarray*}&&L={NH}/| f| ,\end{eqnarray*}$$

where N is the Brunt–Vaisala frequency, H is the scale height, and f is the Coriolis parameter. The Coriolis parameter, f, is 9.1 × 10⁻⁵ s⁻¹ at 15° N planetographic latitude. The pressure scale height, H = k_B T/m g, is 25 km, using a temperature of 160 K (near the NH₃ cloud base), an effective gravity of 22.88 m s⁻² for 15° latitude, and a mean molecular mass of 2.30 Da assuming Galileo Probe composition (Niemann et al. 1998; Mahaffy et al. 2000; Wong et al. 2004). The Brunt–Vaisala frequency, N, describes the static stability of the atmosphere, and has been constrained by a number of methods in Jupiter's atmosphere, including models of cloud formation and vortex circulation, and temperature profiles measured by radio occultations, infrared spectroscopy, and in situ measurement by the Galileo Probe; typical values of N range from 2 to 12 × 10⁻³ s⁻¹ in and between the cloud layers, to 1.6 and to 2.3 × 10⁻² s⁻¹ at 250 mbar just below the tropopause (Wong 2011; Watkins & Cho 2013 and references therein). The atmospheric deformation radius near the NH₃ cloud base is then around 1400 km if cloud condensation creates stability, or 500 km in the free air below this cloud level. The wind jet spans 45 of latitude or ∼5000 km, just broad enough to meet the minimum 3–10 L criterion for possible energy transport from eddies to jets (Conrath et al. 1981). The waves span 2000–3000 km, and are thus consistent with the Simmons (1974) model of the meridional scale of baroclinic instabilities.

Additionally, on Earth, baroclinic instabilities tilt westward and northward with altitude (e.g., Lim & Wallace 1991; Yin & Battisti 2004). The Hubble images at 889 and 275 nm are more sensitive to higher altitudes, but both have little high contrast detail and also show underlying cloud structure, so any such tilts are not obvious, see Figure 7. However, some of the wave crests do have a slight westward tilt with the zonal wind pattern, Figure 5, while others near the cyclone at 202° W are curved. This is consistent with theoretical models (Holton 1992) and Earth observations (e.g., Blackmon et al. 1984).

Zoom In Zoom Out Reset image size