Ground-Based Observations of the Aurora, Ionosphere and Upper Atmosphere of Uranus

Home , Ionosphere

H. Melin, T. S. Stallard, S. Miller, L. M. Trafton, J. McGuire

University of Leicester Space Environment Technologies Overview

• The overarching question: the upper atmospheres of the giant planets are hot! Why?

• Observing Uranus from the ground - attempts to detect aurora. How is H3+ formed?

• Long term monitoring of Uranus’ ionosphere

• Short term and spatial variability of the ionosphere.

• Neptune! Temperature of the upper atmosphere of the giant planets

GIANT PLANET IONOSPHERES AND THERMOSPHERES 335

TABLE III Comparison of predicted and measured exospheric temperatures.

Jupiter Saturn Uranus Neptune

Heliocentric distance (AU) 5.20 9.57 19.19 30.07 Absorbed solar ﬂux (W m−2) 3.7 × 10−5 1.1 × 10−5 2.7 × 10−6 1.1 × 10−6 Texo (observed) [K] 940 420 800 600 Texo (calculated) [K] 203 177 138 132 !Texo (obs-calc) [K] 737 243 662 468

From Yelle et al. 2004 the solar wind to sweep open field lines down the magnetotail, and prevent them rotating with the planet. At the edge of the polar cap, which extends to a colatitude of ≈ 15◦, the ion wind is ≈ 1.7 km s−1. With Saturn’s auroral magnetic field being ≈ 6.5 × 10−5 Tesla (Cowley et al., 2004), Equation (3) gives the field strength ≈ 0.1 V m−1. Cowley et al. (2004) also relate the measured ion angular velocity to the solar wind velocity and the effective Pedersen conductivity: ∗ ∗ !ion = !Sµ0"P VSW/ 1 + µ0"P VSW (6) ! " where µ0 is the permeability of free space, and VSW is the solar wind velocity. This important relationship, first derived by Isbell et al. (1984), holds out the prospect of correlating the measured ion velocities with Cassini measurements of the solar wind velocity, and thereby measuring conductivities, which can be modelled to derive particle precipitation fluxes. Alternatively, in the absence of available space- craft data, the measured values of !ion may be used, with modelled conductivities, to obtain values of VSW.

6. Energy Considerations

Yelle and Miller (2004) have recently compared the measured exospheric temperatures of the giant planets with those calculated from solar EUV inputs alone. Globally, the solar EUV absorbed at Jupiter is ≈ 2.4 TW, while at Saturn it is ≈ 0.5 TW. Table III shows that considerable additional energy sources are required to produce the observed temperatures. Particle precipitation in Jupiter’s auroral/polar regions is estimated to provide an additional 10 to 100 TW (Clarke et al., 1987), a considerable increase on the solar EUV input, although a large fraction of that may be deposited below the homopause, from where much of the UV auroral radiation emanates; below the homopause – as already noted – hydrocarbons radiate away the energy very efﬁciently (Drossart et al., 1993). That means that the actual direct energy input into the upper atmosphere (above the homopause) is probably less than 10 TW globally. Grodent Ways to heat a thermosphere

• Heating by solar EUV radiation. • Heating form lower altitudes via the breaking of gravity waves (e.g. Matcheva et al., 2001 for Jupiter) • Heating injected by the magnetsphere/ ionosphere/thermosphere interaction - Joule heating combined with global re- distribution of this energy (e.g. Melin et al., 2006 for Jupiter). • Other? + The molecular ion H3 H+ +H H+ +H 2 2 ! 3

H2 is ionised by: Solar EUV ionisation Particle precipitation

H3+ traces the injection of energy into the upper atmosphere

Emission from H3+ is observable in the infrared Uranus at equinox P = 17.24± 0.01 h (Desch et al., 1986)

Herbert (2009) - 1986 Voyager 2 UVS observations Uranus at equinox P = 17.24± 0.01 h (Desch et al., 1986)

Strategy: Observe auroral emission as it passes underneath the slit, modulating the + H3 intensity.

Spectrograph slit Data products: - Thermospheric temperatures + - H3 densities + - H3 cooling rates

Herbert (2009) - 1986 Voyager 2 UVS observations Re-detection of H aurora L07105 LAMY ET AL.: EARTH-BASED DETECTION OF URANUS2’ AURORAE L07105

Figure 3. HST/STIS images acquired on (a) 16 Nov. 2011 15:32:10 UT, (b) 29 Nov. 2011 02:09:24 UT and (c) 29 Jul. 1998 06:07:43 UT, with the clear MAMA filter (spanning H2 bands and HHST Ly-a) over - Lamy 1000 s. (top) et Raw al., images. 2012 (bottom) Same as Figure 3 (top) but after subtraction of a background of reflected sunlight and smoothing over 5 pixels (see auxiliary material). White arrows mark features of interest above the noise level: localized bright spots (Figures 3a and 3b), and northern/southern ring-like emissions (Figure 3c). Corresponding planetary configurations are shown with grids of planetocentric coordinates, where blue and red dashed lines indicating the latitude of the northern and southern magnetic poles.

3 to 5 standard deviations (s) above the background level by 120 longitude, which is less than the expected 180 are visible in Figures 3a and 3b (see auxiliary material). longitude difference. These results may be biased by the Using the above disc brightness, the 3s level yields 1–2 kR. incomplete view of both ovals and/or their non-elliptical Despite ACS being more sensitive than STIS, it did not shape. No interplanetary shock was predicted close to obtain comparable results. This could arise from either the observing dates. The time-tag mode revealed that the inten- timing of the ACS measurements and/or the STIS ability to sity of both ovals again varied over timescales of minutes. be more solar-blind, offering better contrast. Interestingly, The N oval for instance intensified during the central 6 min the bright spots in Figures 3a and 3b both lie at the same of the 17 min acquisition time. Signs of sporadic polar longitude (arbitrary) and latitude (between 5 and 15). activity were also observed within both ovals. This preferentially matches the visibility regionÀ of the NÀ oval. Fainter signals of possible auroral origin were also observed 4. Discussion around both magnetic poles and at longitudes consistent with the planetary rotation in two thirds of the STIS images. Orbits [13] Aurorae observed in 1986, 1998 and 2011 sample a 17 and 9 lay 1 and 2.5 days after the pressure fronts C and A, quarter of a Uranus revolution around the Sun. Their char- respectively, but some northern fainter signal was already acteristics provide insights on how diverse solar wind/ observed during orbit 7, 2 days prior to orbit 9. Using the magnetosphere configurations influence their mutual inter- STIS time-tag ability, we found that auroral signals in action and affect auroral processes. Figures 2a and 2b vary within the 1000 s duration of each [14] In 1986, the large tilt between the magnetic and image. The two hot spots mainly brighten during the first and rotation axis led the N and S magnetic poles rotating second half of the two image intervals, respectively. uniquely within the dayside and nightside hemispheres respectively, a unique situation in the solar system where [12] We have also re-analyzed previous HST observations of Uranus. In July 1998, the planet was observed twice with the two dominant regimes governing magnetospheric STIS over 1.5 days. The image displayed in Figure 3c plasma transport (solar wind convection and planetary reveals two roughly continuous, distinct, rings of emission. rotation) act in perpendicular planes and are therefore While their intensity per pixel lies between 1 and 3s above decoupled. This gives rise to a significant helical magne- the noise level, the signal integrated along model ovals, fit- totail [Behannon et al., 1987]. The prominence of night- ted to their visible part with partial ellipses, reaches 4 to 5s side emissions, consistent with this configuration, led to (see auxiliary material for details). N and S fitted emissions consider auroral precipitation as driven by solar wind convection through nightside reconnection processes, as for extend at latitudes from 5 to +70 and from 15 to +30, respectively, the N ovalÀ being larger than theÀ S one, as the Earth [Herbert, 2009]. These emissions varied on few- expected. Both ovals are centered at around +40 and 11 day timescales [Broadfoot et al., 1986], in agreement with À typical convection-driven plasma residence times of 1– latitude respectively, both similarly shifted by +25north of the latitude of their associated magnetic pole, and separated 3 days [Belcher et al., 1991]. Interestingly, we note that the

4 of 6 Long-term cooling of Uranus’ ionosphere New 2012/2013 observations 1. Bi-annual variability Geometric season - heating from below

Phase delay Geometric Cooler Equinox season

Hot Solstice Hot Solstice

Cooler Equinox 2. Annual variability Magnetic season - heating by MIT interaction Magnetospheric Tepid Equinox season

Hot Solstice

Cold Solstice

Tepid Equinox What drives the heating of Uranus’ thermosphere? Magnetic/geometric season?

? What drives the heating of Uranus’ thermosphere? Magnetic/geometric season?

? 2006 Keck Observations

• Slit aligned with rotational pole of the axis. • Observed emission as Uranus rotates underneath the slit • Noon meridian is always bright in H3+. • Variability across the slit - aurora? VLT CRIRES 2013

Dawn Dusk

South 642 H. MelinA et al.word on Neptune ﬂux density at Uranus, auroral processes are still likely only to be

responsible for a small fraction of the total H3+ density. The model of Lyons (1995) predicts a H3+ peak density of 3 250 cm− at an altitude of about 1350 km above the 1 bar 13 2 level, giving a column-integrated density of 4–5 10 m− (Feuchtgruber2006 & EncrenazKeck 2003). NRISPEC× Here, we present the analysis of Keck II NISPEC observations of Uranusobservations and Neptune in an effort of to improve Uranus the upper limit for the column-integrated H+ density of Neptune. and 3 Neptune 2DATA&ANALYSIS Spectra were obtained using NIRSPEC on Keck II on 2006 Septem- ber 5, at wavelengths sensitive to emission from the H+ molecular 3 New limits on H3+ abundance on Neptune 643 Neptuneion. NIRSPEC (McLean et al. 1998) is a medium to high resolution spectrograph (R 25 000) mounted on the Keck II 10-m telescope Figure 1. On the right-hand side is a K-band image of Neptune showing on the summit of= Mauna Kea, Hawaii. We used the KL grating the position of the NIRSPEC spectrograph slit. On the left-hand side is the (2.134–4.228 µm) with the 0.288 24 high-resolution slit giving a continuum spatial profile of Neptune derived from both the K-band image 1× profile and the L′ spectral profile (summed over all three spectral orders). pixel scale of 0.144 arcsec pixel− . This produces a cross-dispersed The K-band image profile has been smoothed to match the seeing of the L spectrum with five spectral orders within the L atmospheric win- ′ ′ observations. dow – a region that contains both the brightest rovibrational H3+ emission lines and a region of methane absorption in the under- on 2006 September 4 (24 h prior to the L observations), utilizing lying atmosphere. Each of these orders contains some degree of ′ adaptive optics (AO). The pixel scale is 0.01 arcsec and the total Figure 3. curvature,The regions in where both the H+ spectralemission and is found spatial at dimension; Uranus (bottom) so prior and to Neptune (top). The lines are shown in order of decreasing wavelength, with the Uranus 3 integration time is 6 min. The image shows two distinct bands of longest wavelengthany reduction, at the each left. individual The far right-hand spectral image shows was straightened the co-add us-of all the lines, with no emission detected for Neptune. All higher order R regions reflected sunlight, with the northern being brighter than the south- are adjacenting to the strongREDSPEC telluric IDL absorption.package. This The paper intensity only scale considers for+ each the image orders is the same for each planet. No H3 seen onern bandNeptune: by a factor of 2. This bimodal cloud structure was also that contain strong H+ emission: order 22 (3.40–3.46 µm), order 21 3 observed by Feuchtgruber∼ & Encrenaz (2003; in K and L bands) (3.56–3.62 µm) and order 19 (3.94–4.01 µm). ′ from a ionospheric point of viewand by Max - et Uranus al. (2003; in H and ≠J bands),Neptune but they both noted a Table 1 lists the observations of both Uranus and Neptune anal- wavelengths. Note that in order 22, and to a lesser extent in order brighter reflection from the Southern hemisphere. ysed here. The spectrograph slit was aligned along each planet’s 21, there is significant absorption of the emission by water in the The intensity profile of the reflected sunlight across the planet, rotational axis. There were a total of 64 exposures of Neptune, each Earth’s atmosphere. This absorption is difficult to correctFeuchtgruber for, and it &along Encrenaz the direction of (2003 the slit, can), beMelin seen on et the left-handal. (2011) side 60 s long, contained within 16 ABBA sequences, where B is a beam produces narrow regions with large noise, seen in both the Uranus of Fig. 1. The profile derived from the K-band image has been offset from A such that the target remains on the slit for both expo- and Neptune spectra. smoothed by 0.7 arcsec (70 NIRC2 pixels) to match the seeing sures. The total set of observations includes a larger set of Uranus of the L NIRSPEC spectrum, which does not utilize AO. The Orderspectra, 19 is not but affected in order by to compare terrestrial equivalent water vapour. levels of The signal-to-noise Uranus ′ Neptunian longitude difference between the two profiles is 180◦ spectrumratio (top only right-hand 64 of these side were of Fig. used, 2) giving was fitted a total with integration a model time H3+ for ( 24 h) suggesting that the Northern hemisphere is consistently∼ spectrumUranus assuming and Neptune condition of of 64 local min each. thermodynamic equilibrium brighter∼ over a wide range of longitudes during this epoch – there is (LTE). WeThe derive spectral a column-averaged image for each target rotational was dark H+ temperature current subtracted, of 3 good correspondence between the distribution of reflected sunlight 604 8Kflat-fielded, and a column-integrated co-added and flux calibrated density using of 1. a13 240-s0 exposure.07 of in the K-band image and L spectrum. 15 ± 2HD 218639, an A0V star with T 10 000 K and± a K magnitude× ′ 10 m− . This is well within the rangeeff of the previously reported The co-added spectrum of Uranus and Neptune can be seen in of 6.369 (Cutri et al. 2003). The seeing= during these and subsequent values, e.g. Trafton et al. (1999) and Encrenaz et al. (2003), con- Fig. 2 – Uranus shows distinct H+ emission lines across all three observations was 0.7 arcsec . 3 firming that the flux calibration is sensible. The small errors of orders, whereas Neptune has no discernible emission at the same this fit highlightThe angular the excellent diameter signal-to-noise of Uranus on the ratio sky achievable at the time with of observation was 3.6 arcsec, covering 26 pixels, whereas the angular NIRSPEC when observing H+ emission from the gas giants. diameter of Neptune was3 2.3 arcsec, covering 16 pixels. The spatial Figure 4. The spectral profile of the sum of the regions where H emission For Jupiter, the Q branch emission at 3.9 µm is the least affected 3+ position of Uranus was easily located in the spectral image by its is observed for Uranus, seen in Fig. 3. No emission is detected for Neptune. by departures from LTE (Melin et al. 2005). However, the modelled prominent Q and R branch H3+ emission lines – these were also used reductionto in establish line emission the wavelength intensity scale ranges for each between order. 70 per cent for the Q(1, 0−Since) line and H3+ emission 4 per cent lines for the were R(6, absent 6+) ofin the the LTE co-added intensity. spectral Since the width of the summed H3+ line profile of Uranus in Fig. 4 On Neptune,image of assuming Neptune, the the H position2 density subtended and temperature by the planet values on of the results mainly from instrumental broadening, we can assume that spectrometer slit was deduced from the continuum of sunlight re- Broadfoot et al. (1989) at the H3+ peak at 1350 km (Lyons 1995), the the emission line profile from Neptune would have the same width. flected from the reflected sunlight continuum from the disc of the H2 collisional population lifetime for the H3+ ν2 level is 0.5 ms (us- Using the standard deviation of the Neptune intensity in Fig. 4, i.e. planet. Fig. 1 shows a K-band image of Neptune∼ (right-hand side) ing the proton-hopping rates of Theard & Huntress 1974), compared the noise, as the maximum value for the H3+ intensity, we derive a taken with the Keck II Near InfraRed Camera (NIRC2) at 05:35 UT 2 1 to the radiation lifetime of 8 ms (Neale, Miller & Tennyson 1996). maximum flux of 0.9 µWm− sr− for the 13 spectral lines. This suggests that the fundamental rovibrational energy levels of Table 1. The Keck II NIRSPEC data. !ti is the total Feuchtgruber & Encrenaz (2003) used a temperature of H3+ have enoughintegration time to time kinetically for each target. equilibrate with the ambient 550 100 K to calculate the maximum column density of H3+ to ± 7.1 14 2 temperature and that the assumption of LTE is valid, and there- be 2.9 +1.8 10 m− . This temperature was obtained by as- Target UT date UT time start !ti (min) − × fore the H3+ flux from Neptune will not be significantly reduced by suming the altitude of the peak H+ density as 1350 km above the ! " 3 ∼ non-LTE effects.Uranus 2006-09-05 07:26 64 1Figure bar level 2. The (Lyons spectra 1995), of Uranus referenced (top) and Neptune against (bottom) the Voyager for the three 2 tem- NIRSPEC orders considered here. The Neptune spectra has been offset by Fig. 3 showsNeptune each of 2006-09-05 the 13 identified 05:27 H3+ emission 64 lines in perature profile of Broadfoot et al. (1989). Using this temperature 0.3 intensity units for clarity. the Uranus spectral image – these lines are contained within 12 for− the line flux obtained here, we derive a maximum column- 4.8 13 2 wavelength regions. integrated H3+ density of 1.5(+0.9) 10 m− . Using the Broadfoot C 2010 The Authors. Journal compilation− C ×2010 RAS, MNRAS 410, 641–644 The line-of-sight orbital velocity difference between Uranus and ⃝ et al. (1989) temperature of 750⃝ 100 K, we derive a column den- 2.8 12 2 ± Neptune, as viewed from Earth at the time of observation, is sity of 2.8(+1.1) 10 m− . More generally, the upper limit for 1 − × 2kms− , which is well below the velocity resolution of NIRSPEC the H3+ column density as a function of temperature can be seen in 1 of 12 km s− .Consequently,weexpectanyH3+ emission lines at Fig. 5. Neptune to be found at the same spectral position in the spectral ar- ray as for Uranus. Fig. 3 shows the equivalent regions for Neptune, with no obvious line emission present. In order to achieve the maxi- 3RESULTS&DISCUSSION mum possible signal-to-noise ratio, these regions were summed for each planet, shown on the right-hand side of Fig. 3. These sums 3.1 Variability of hydrocarbon absorption were integrated in the spatial direction, the result of which is seen In these observations, taken in 2006 September, the solar reflection in Fig. 4. H+ emission remains undetected from Neptune. 3 component from the Northern hemisphere of Neptune is observed Using the H+ line list of Neale et al. (1996), one can calculate the 3 to be brighter than the Southern by a factor of 2. Observations total emission per molecule for the sum of the 13 emission lines that ∼ 13 obtained in 1999 in H and J bands (Max et al. 2003) and observations are added here. The intensity per molecule, Imol (T), as a function of temperature, T, takes the approximate form taken in 2002 in K and L′ bands (Feuchtgruber & Encrenaz 2003) both see a more intense solar reflection in the Southern hemisphere. 13 3 2 18 The time-scale is too short to suggest seasonal variability (P Imol(T ) (30.4 0.17T 0.23 10− T ) 10− (1) = − + × × 165 yr), but it does tell us that the hydrocarbon layer is very dynamic,= 1 in units of W sr− . which was also noted by Feuchtgruber & Encrenaz (2003).

C C ⃝ 2010 The Authors. Journal compilation ⃝ 2010 RAS, MNRAS 410, 641–644 Summary

• The cooling of Uranus’ upper atmosphere continues!

• Long term monitoring will shed light on the mechanism that is present on all the giant planets.

• Observations of H3+ may show weak auroral emissions - limited by S/N.

• Uranus and Nepune are, ionopherically speaking, are very different.