Steven Miller et al- Mid-to-Low Latitude H3^+ Emission from Jupiter

8/3/2019 Steven Miller et al- Mid-to-Low Latitude H3^+ Emission from Jupiter

1/11

ICARUS 130, 5767 (1997)ARTICLE NO. IS975813

Mid-to-Low Latitude H3 Emission from Jupiter

Steven Miller and Nicholas Achilleos

Department of Physics and Astronomy, University College London, Gower Street, London WC1E 6BT, United Kingdom

E-mail: [email protected]

Gilda E. Ballester

Department of Atmospheric, Oceanic and Space Sciences, University of Michigan, 2455 Hayward, Ann Arbor, Michigan 48109

Hoanh An Lam and Jonathan Tennyson

Department of Physics and Astronomy, University College London, Gower Street, London WC1E 6BT, United Kingdom

Thomas R. Geballe

Joint Astronomy Centre, 660 N Aohoku Place, Hilo, Hawaii 96720

and

Laurence M. Trafton

Department of Astronomy and McDonald Observatory, University of Texas at Austin, Austin, Texas 78712

Received April 21, 1997; revised July 9, 1997

by workers modeling the atmospheres of the outer pl(see, e.g., Atreya and Donahue, 1976). The VoyagWe present measurements of the mid-to-low latitude H3

emission from Jupiter, derived from a spectroscopic study of the spacecraft detected an ion with a mass:charge raplanet carried out on the United Kingdom Infrared Telescope 3 : 1 in Jupiters magnetosphere during its flyby o(UKIRT) on Mauna Kea, Hawaii, on May 35, 1993. The planet in 1979 (Hamilton et al., 1980). Since its spemeasurements indicate ionospheric H3 temperatures 800 K scopic detection in the jovian aurorae in 1988 (Droand column densities of the order of 1011 cm2. The emission et al., 1989), H3 emission has been used to probe aulevels depend strongly on latitude and longitude, but are gener-

activity and morphology using both spectroscopically of the order of 101 erg s1 cm2, indicating that the cooling

imaging techniques (see review by Miller et al., effect of H3 is a significant factor in the ionosphere. Theseand references therein). In recent years a numbemission levels also strongly suggest either that aurorally pro-infrared spectroscopic studies of Jupiter have shownduced H3 is being transported to nonauroral latitudes or thatemission from the fundamental rovibrational bansources in addition to solar EUV are required to produce thethe H3 molecular ion may be detected from all ionisation and excitation energy necessary to account for the

observed H

3 emission. This view is supported by comparing of the planetary disk (Miller et al., 1992; de Berghthe emission profiles as a function of latitude with those ob- 1992; Ballester et al., 1994, henceforth Paper I).tained from a jovian global circulation model which has auroral gave rise to the expectation that H3 emission couelectron precipitation and solar EUV as ionisation inputs. The used to probe and monitor conditions in the nonauspatial distribution of H3 emission suggests that this ion may ionosphere. Paper I, for instance, considered a potbe a useful probe of Jupiters magnetic field at subauroral

link between H3 emission and the Lyman- Bulatitudes. 1997 Academic Pressregion around Jupiters equator which shows enhaLyman- emission. Also, during the collision of CShoemakerLevy 9 with Jupiter in July 1994, se1. INTRODUCTIONgroups studied the effects of the impacts on the sphere by monitoring this ion (Dinelli et al., 1997The importance of the H3 molecular ion as a constit-

uent of the jovian ionosphere has long been recognized et al., 1996; Schultz et al., 1995).

57


2/11

58 MILLER ET AL.

TABLE 1Recently, we have published results from a study ofUKIRT 3.47-m Data Obtained during 1993jovian H3 emission carried out on the United Kingdom

Infrared Telescope (UKIRT) on Mauna Kea, Hawaii, onDate Time (U.T.) CML (System III)

May 35, 1993 (Lam et al., 1997, henceforth Paper II).While this paper concentrated on the auroral H3 emission, May 3 05:36 102

07:42 180some details of nonauroral emission were also included.08:32 210In general, Paper II explained that H3 emission is produced09:46 254by the radiative relaxation of collisionally excited ions,11:37 324which exist in a state of quasi-LTE. The emission, espe- May 4 05:14 241

cially in the mid-to-low latitude regions, is generally opti- 07:33 32508:55 40cally thin. Paper II also introduced a parameter E(H3 ),

May 5 05:42 47which allowed for the strong coupling between fitted tem-09:11 174peratures and column densities. This parameter is defined

by calculating the total emission per molecule at the appro-priate fitted temperature, assuming local thermal equilib-rium, and multiplying it by the corresponding fitted column dinal variations, while eliminating local time effectdensity. (For further discussion of the reliability of this discussed in Paper II, temporal variations also play aparameter, see Paper II.) E(H3 ) is particularly useful for in determining the H3 emission spectrum. But compathe subauroral latitude study, where emission levels are of Tables 2 (CML 324, May 3, 1993) and 3 (CML

relatively weak and spectra rather noisy, and where we May 5, 1993) of Paper II shows that this is not too are forced to use a rather assorted selection of H3 spectral a problem to prohibit using data obtained at differentlines for the purposes of fitting. to build up a picture of the spatial variation of the j

In this paper, we examine the behavior of subauroral H3 emission. This is particularly so in the mid-to-lowH3 in terms of temperature and column density variations, tudes discussed here, where temperatures at equivthe resulting planetary distribution of E(H3 ), and what locations throughout the region vary within the statethis means in terms of the energy balance in the ionosphere. rors on the two days, and the more significant E(H3We compare the observed H3 variation with latitude with files are within 10% of each other at all non-aurorathat predicted by a jovian global circulation model (JIM, tudes.Achilleos et al., 1997).

2. DATA ACQUISITION AND ANALYSIS TABLE 2Comparison of 1992 and 1993 Data Obtained

All data were obtained with UKIRTs long slit grating at CML 102spectrometer; full details of the UKIRT data acquisition

1992 data 1993 dataand reduction have already been given in Papers I and II,and only the essential details are recapped here. This pres-

Position N(H3 ) Position Nent study is based solely on the CGS4 spectra obtained in (Lat/km) T(K) (1012 cm2) (Lat/km) T(K) (1012the wavelength range 3.373.57 m (central wavelength

8000a 819 0.13 10,484a 1053 03.47 m) and with the slit aligned northsouth along the3000a 869 0.81 183a 822 5jovian central meridian longitude (CML). Times and cen-N 67 706 6.71 N 56 639 3tral longitudes observed are given in Table 1. Each spec-N 48 664 1.78 N 42 763 0

trum consisted of 2.4 min of exposure time on Jupiter and N 36 983 0.19 N 30 823 0

2.4 min on adjacent sky (to remove the sky background N 25 1043 0.14 N 21 761 0N 15 1250 0.08 N 12 976 0from Jupiter) and took6 min to complete (correspondingN 06 1100 0.10 N 04 761 0to a jovian rotation of 3.6). The pixel size was 3.08 S 01 1140 0.10 S 04 813 03.08, corresponding to 10,000 10,000 km on a planeS 09 829 0.28 S 12 764 0

at the distance to Jupiter. The pixel thus subtended 8.0 S 18 1230 0.07 S 21 741 0of longitude 8.2 of latitude at the sub-Earth point, S 28 1140 0.10 S 30 752 0

S 39 863 0.22 S 42 803 0rising to 12.5 of longitude 13.1 of latitude at a sub-S 51 827 1.29 S 56 943 0Earth point of 50 along the CML. Since all the dataS 75 900 5.00 183a 786 4used in Papers I and II and this paper were obtained with3000a 680 11.54 10,484a 764 0

the spectrometer slit aligned along the CML, they referto the planetary local noon. This observing strategy was a Distances in kilometers above the limb of the planet, determi

fitting program (see Paper II).adopted in order to concentrate on longitudinal and latitu-


3/11

MID-TO-LOW LATITUDE JOVIAN H3 EMISSIONS

Details of the method of analysis of the data have also the H3 line positions, giving an estimate of the continemission of accuracy appropriate to the signal-to-noilargely been given in Papers I and II. There is, however,

an important difference between Paper I and the current tio and the temperature/column density fitting proceThe effect of this improved background estimation ostudy in the handling of the background continuum emis-

sion. At 3.47 m, much of the solar infrared radiation 1992 data is discussed in the next section.incident at Jupiter is absorbed by the 3 band of methane,present in the stratosphere with a mixing ratio 103. To

3. H

3 TEMPERATURE AND COLUMN DENSITa first approximation, the absorbing column is enhancedalong the line of sight by a factor of 1/cos(), where isthe local solar zenith angle (equivalent to the latitude, Typical fitted H3 temperatures and column densiti

given in Tables 2 and 3 of Paper II. They show that ousince our observations were made with the slit alignedalong the CML). As a result, in the auroral regionsand of the bright auroral regions, derived temperatures ra

from below 700 K (as low as 639 K in one instancat higher latitudes generallyvery little background radia-tion is seen in moderate resolution spectra. Since the H3 around 850 K (up to 875 K, in another), with 2 e

between 10 and 15%. Column densities prior toemission is generally much greater (typically up to 50times) than the continuum, the effect of the background of-sight correction ranged from values near the eq

around 1 1011 cm2 to up to six times that valon fitted temperatures and column densities is negligible.But at latitudes close to the equator, the intensity of the the higher latitudes (50), with errors of 55%

equator falling to around 20% at higher latitudes.continuum reflected solar IR is greater, since the absorbing

methane column is less than towards the poles, and the In Table 2 we compare the results obtained in Pawhich is based on 1992 data and fitted with a conH3 emission up to 50 times weaker than in the auroral

regions. As a result, the continuum may be up to 25% background, with those derived in the present study, wwe also made measurements at CML 102 in 1993of the line intensity. Clearly then, the treatment of this

continuum in fitting the mid-to-low-latitude spectra may fitted them with a background produced as descabove. Paper I reported that there was a temperatursignificantly affect the values for temperature and column

density derived. Paper I assumed that this could be approx- hancement of200300 K in the fitted equatorial temtures of the 1992 dataset compared with that deriveimated by a constant across the entire wavelength region

(0.25 m). During the course of this study, however it the auroral regions. Dividing the planet as a wholethree regions to improve the statistics of the data (pabecame clear that this treatment, adopted for the 1992 data

because of their rather poor signal-to-noise level, was an larly in the central latitudes), they found a temperatu1220 120 K in the central region (latitudes 30oversimplification.

The background in the 3.47 m spectral window could compared with 813 41 K in the southern higher latregion and 734 37 K in the northern. Among have structure if there were absorption features in the

incident solar radiation or as a result of differential absorp- things, the authors of Paper I considered that their rcould provide support for a new explanation fotion by jovian stratospheric methane; Drossart et al. (1996)

have recently ascribed one feature at 3.52 m (see Fig. 2, Lyman- Bulge proposed by Ben Jaffel et al. (1992later extensively modeled by Sommeria et al. (1995). TPaper II) to the latter effect. Solar absorption features

are at the level of a few percent of the solar continuum models suggested enhanced temperatures as a resstrong winds flowing from the auroral regions.emissionand therefore about 1% of the peak intensities

of our spectra, when scaled to the background level (maxi- The1993 results, however, indicate less spatial variain temperature than was suggested by Paper I. Wemum 25% of the line intensity), well within the noise of

our datathroughout the spectral range of interest, with consider that the higher central region temperature

tained in that paper were an artifact which resultedthe exception of one unidentified line at 3.523 m andlines due to silicon and magnesium between 3.396 and an inadequate treatment of the continuum backgr

radiation, particularly at shorter wavelengths. This res3.400 m (Livingstone and Wallace, 1994). Fortunately,none of these coincide with any of the strong H 3 emission in the higher energy H

3 lines having their intensity

cially enhanced, hence apparently increasing the fittedfeatures used in fitting the data presented here. For simplic-ity, we did not use a line-by-line calculation of the absorp- perature. The 1992 data are noisier that that of 1993,

ever, and reliably fixing the background was conseqution due to jovian stratospheric methane, since our primaryconcern in this study was the emission due to H3 . In the more difficult and the fit less reliable. This mean

reanalysis of individual rows, using backgrounds dpresent study, therefore, the H3 lines were first manuallyremoved from the spectrum and the resulting background mined in the same way as for the 1993 data, gave unce

ties too large to make the results useful. But it was pobetween the lines was fitted by a low-order polynomial.This background was then smoothly extrapolated under to reanalyze combined rows of the central region


4/11

60 MILLER ET AL.

FIG. 1. Global jovian H3 temperature distribution.

produced a mid-to-low-latitude temperature for the central 2 or below allowed too much flexibility in the plottinproduced interpolation artifacts; stiffnesses of 79 weregion (latitudes 3033) of 930(150/110) K, con-

siderably lower than 1220 (120) K reported in Paper I. large as to completely override genuine spatial variatThe maps were produced from the full 1993 dataset, wNonetheless, as will be seen below, it does seem as if there

is some coincidence between the pattern of H3 emission includes data taken at a central wavelength of 4.00However, for reasons explained in Paper II, the 4 mand the location of the Lyman- Bulge. Although there

may be a physical connection between the phenomena, only cover the auroral regions. (The full dataset i

useful for mid-to-low-latitude regions since the periauthis coincidence may be also fortuitous.In Figs. 1 and 2, the fitted results for temperature and values derived at 4 m help constrain the overall plo

fill in some of the gaps.) As a result, there are somecolumn density, corrected for line-of-sight effects, are pre-sented as a function of planetary location. (As reported in in the data at mid-to-low latitudes, particularly bet

System III longitudes III 102 and 174, and betPaper II, at the mid-to-low latitudes discussed here, theline-of-sight correction was approximated by a secant ef- III 325 and III 40. The spatial resolution o

pixels, given in Section 2, and the smearing due tfect.) Contours between the datapoints were interpolatedusing a two-dimensional spline interpolation with the Un- 3.6 planetary rotation during each 6-min exposure,

also be taken into account. Thus some of the small-iras graphics package. A moderate stiffness of 4 (stiffnessesanywhere between 3 and 5 made little difference to the mid-to-low-latitude detail produced in the figures ma

be a product of program interpolation between the overall plot) was chosen to trace real point-to-point varia-tions without introducing spurious features. (A stiffness of points. Outside of the auroral regions, therefore, the


5/11


FIG. 2. Global jovian H3 column density distribution in units of 1012 cm2. (Corrections for line of sight effects have been made.)

should be used to derive trends rather than indicate specific 4. H3 TOTAL EMISSION, E(H3 )local knowledge.

As explained in Paper II, the derived column denIn the mid-to-low-latitudes, our data show that there isand temperatures are strongly anticorrelated, and tha tendency for the temperature to fall from around 850ues resulting from our fitting procedure may cover a950 K, close to the auroral regions, to between 700 andrange of parameter space (see Fig. 6 of Paper II), g800 K in regions centered on N 40 and S 35. But towardrise to large errors in the individual parameters. As a rthe equator, the temperature does appear to climb, once

a new parameter E(H

3 ) was introduced. This parammore, closer to the near-auroral 850950 K range. In addi-gives a value for the total emission due to the H 3tion, there appears to be a warm channel linking the north-derived by combining the fitted temperatures and coern and southern auroral regions centered around III densities, and assuming LTE. Paper II showed tha170210, a regionreasonably well sampled by our dataset.parameter has associated errors of 20% to 25% foThis channel could possibly extend to lower longitudesmid-to-low-latitude regions.where we do not have coverage in the mid-to-low-latitude

E(H3 ) is a useful parameter as it gives the coolingregion, although the mapping is somewhat constrained byfor this important constituent of the jovian ionospauroral and periauroral 4 m data at III 160 (Table 1Outside the auroral regions, the value ofE(H3 ) (afterof Paper II). The column density map shows rather lessof-sight correction) varies between 0.2 and 0.5 erlatitudinal structure than the temperature map. Densitiescm2 at latitudes around 40, down to 0.05 to 0.10 erfall from a maximum of 1.0 1012 cm2 at latitudes

around 55 to below 0.1 1012 cm2 at the equator. cm2 around the equator. These levels are higher tha


6/11

62 MILLER ET AL.

the model does not generate supersonic auroral wind(iii) no particle precipitation outside of the auroral ris included. With these restrictions the model predictat latitudes immediately adjacent to the auroral regwhere only the effect of solar EUV is present, themission (before line-of-sight correction) should falidly toward its minimum value at the equator. This ihowever, what is observed in the data. Instead, the obtions show that E(H3 ) decreases much more slowlyfunction of decreasing latitude, and that the modelously underestimates the intensity close to the aurorgions. We shall also return to this point later.

In Fig. 4, we present mid-to-low latitude details oE(H3 ) map (corrected for line-or-sight effects) presas Fig. 7 of Paper II. The figure indicates that, supposed on the general tendency for E(H3 ) to fall witcreasing latitude, there are considerable longitudinal

FIG. 3. Comparison of jovian H3 emission, E(H3 ), profile at III tions. The first point to note is that the minimum i

40 predicted by the JIM global circulation model with that derived from H3 emission does not always coincide with the equatthe 1993 dataset. Squares: data points. Solid curve: JIM profile convolved

particular, there appears to be a region between IIIwith CGS4 spatial resolution. (No line of sight corrections have beenand 170 where the minimum is as much as 20 nomade.)the equator. (The region between III 102 and 174sampled in the mid-to-low latitudes.) Other longituregions have emission minima which differ from the

previously been considered. At the equator the total jovian tional equator by lesser, though not insignificant, amoH3 line flux is comparable to the incident flux of solar It is also noteworthy that, although the present EUV(Lean, 1987)a point we shall return to later. Recent does not support temperatures as high as 1200 K assocmodeling by Waite et al. (1997), making use of the cooling with the Lyman- bulge, insofar as our dataset samprates published here (first reported in Lam, 1995), has this feature does appear to coincide with a region ofshown that inclusion of cooling due to H3 considerably tively elevated temperatures (Fig. 1) and with the ralters the thermal structure of the nonauroral ionosphere. of least H3 emission. This emission minimum results

In Fig. 3, we compare the latitudinal profile of E(H3 ) the lower column densities which accompany the elewith that predicted by a jovian ionospheric model (JIM, temperatures, raising once more the possibility suggAchilleos et al., 1997). The exact details of the model are in Paper I that the emission in this region is being prodnot relevant to this study, except insofar as JIM is a three- by H3 produced at higher altitudes than in the audimensional global circulation model which allows for cou- regions, i.e., closer to the nanobar than the 0.1 micpling between different latitudes and longitudes via self- level, a point to which we return later. The warm chconsistently generated winds. Calculations are carried out linking the northern and southern auroral regions aron a grid of 2 in latitude and 9 in longitude. Currently, III 170210, noted above, also shows up as a rthe energy inputs to the model of importance to this study of brighter emission reaching (almost) right acrosare (i) electron precipitation of 10 keV at 8 ergs s 1 cm2 equator.confined between the footprints of the Io Plasma Torus

and field lines (as predicted by the O6 plus current sheet 5. CANDIDATE EXPLANATIONS FOR THEmodel of Connerney (1993)) which correlate to the magne- OBSERVED H3 EMISSIONtospheric region in which plasma corotation breaks down;(ii) approximately 4 102 erg s1 cm2 of H-ionizing The results outlined above show that the H3 molesolar EUV radiation incident at the subsolar point on the ion is to be found in the mid-to-low-latitude regioplanets atmosphere, with 50% of the total contributing concentrations sufficient for it to radiate away a conto ionospheric heating (consistent with Atreya, 1986); and able fraction of the energy received by Jupiters u(iii) standard jovian ionospheric chemistry (e.g., Strobel atmosphere. But standard discussions (e.g., Atreya, and Atreya, 1983). The current restrictions of the model indicate that there is too little incident solar EUV twhich are relevant to this study are (i) that the auroral count for the observed H3 densities and emission, energy input is more dispersed than the latest imaging ion is being produced and excited in situ. It is ther

important to understand the mechanisms by which would suggest is appropriate (Clarke et al., 1996); thus (ii)


7/11


being transported to or produced in the planetary locations H3 from the aurorae to lower latitudes, the concentris too low to account for the mid-to-low latitude concewhere it is found.tions observed, and where the auroral production rathigh, the lifetimes and wind speeds are too low. The f

5.1. Transport Processesof JIM to reproduce the observed latitudinal distribuusing just auroral particle precipitation, horizontalOne way of producing high H3 concentrations (and thus

emissions) in the mid-to-low latitudes could be to transport sonic winds, and solar EUV, supports this.It is, however, possible that auroral H3 produced it there from the auroral regions where it is produced in

high concentrations. According to Atreya (1986), were the 0.1bar level is being vertically transported to the nalevel, where it is then being dispersed to lower latitauroral energy of Jupiter to be evenly and efficiently spread

across the entire planetary disk, it would amount to 0.4 (At present, vertical winds in the JIM model are comusing hydrostatic equilibrium in the vertical direerg s1 cm2, and would thus be more than enough to

account for the H3 emission in the mid-to-low latitudes. rather than an independent momentum transfer equaSommeria et al.s (1995) model indicates that the tWaite et al. (1983) suggested that this distributive pro-

cess might be achieved by strong thermospheric winds. time from the auroral regions to the equator is less104 s. In the nanobar region, electron densities are typThey showed that diluting the auroral energy in this way

produced temperature profiles in agreement with those 102 cm3 and would allow H3 lifetimes of 104 to

enough time for it to be transported across the ededuced by Voyager.The impact of Comet ShoemakerLevy 9 on Jupiter in planet. Recently, evidence of vertical transport has

observed in auroral Lyman- emissions (PrangeJuly 1994 enabled planetary astronomers to obtain valuesfor a number of important parameters concerned with the 1997). Another contribution to the transport of en

and hence H3 , from the auroral to subauroral regionjovian atmosphere. One of importance to this study is thespeed of meridional winds in the upper atmosphere. Clarke be due to the presence of long-lived H and small am

of He. This latter ion may be formed by particle pret al. (1995) showed that the spread of the impact sitescould be accounted for by winds in the upper stratosphere tation and has a recombination rate with electro

1011 cm3 s1, giving it a lifetime of at least 106 sof between 30 and 50 m s1. This value was shown to beconsistent with the timescale involved with the quenching at the 0.1 bar level. It reacts slowly with H2, w

rate of 1013 cm3 s1, giving rise to products suof the southern aurora (Miller et al., 1995), presumablydue to the transport of material in the lower ionosphere, HeH and H2 , both of which form H

3 (Roberge

Dalgarno, 1982). At the nanobar level, where [H2] and with the 10100 nanobar wind speeds predicted by JIM(Achilleos et al., 1997). While such winds might transport tween 1010 and 1011 cm3, high velocity winds could

port H and He over several thousand kilomenergy efficiently, however, they could not account for theobserved distribution of H3 densities away from the auro- assisting the spread of H

3 to the mid-to-low latitude

ral regions. The dissociative recombination coefficient forthe reaction of H3 with electrons remains controversial 5.2. Joule Heatingbut is now generally thought to be approximately 107

cm3 s1 (Mitchell, 1990, and private communication). At The relatively high level of H3 emission might alexplained in part by the generation of energy as a relectron densities typical of the auroral regions where it

is produced ([e] 104105 cm3 at 0.1 bar), H3 lifetimes of joule heating, due to the frictional effect of ions hortally accelerated by the electric field through the neare therefore between 100 and 1000 s. Thus velocities in

excessof10kms1, rather than the 100m s1 which seems atmosphere. Atreya (1986) estimates that the heightgrated Pedersen conductivity of the nonauroral jprobable in the upper stratosphere/lower ionosphere, are

required to transport the ion the tens of thousands of ionosphere is of the order of 0.2 mho, about 50 timethan its auroral counterpart. He also estimates the kilometers away from the auroral regions to where it is ob-

served. heating produced at auroral latitudes to be about 5s1 cm2. Given the same electric field strength atWind speeds suggested by Sommeria et al. (1995) are,

on the other hand, much greater than thisof the order auroral and mid-to-low latitudes, one might therefopect joule heating in the nonauroral latitudes to be oof 10 km s1. However, these winds are being produced at

high altitudes (in the nanobar region) where models (e.g., order of 0.1 ergs s1 cm2. This is likely, however, somewhat of an overestimate, since higher electricAtreya and Donahue, 1976; McConnell and Majeed, 1991)

predict that H3 is being produced in the auroral regions strengths are expected in the auroral regions than atto-low latitudes. Nonetheless, if this argument is coat much lower concentrations than at the 0.1 microbar

production peak. One is left with the conclusion that, at and if the level of joule heating were close to 0.1 ercm2, this mechanism could account for a fraction oaltitudes where wind speeds may be sufficient to transport


8/11


9/11


observed mid-to-low latitude H3 energy output, depending comparable to the input energy proposed by Atreya (to account for the observed H2 band intensities, as won the efficiency with which it is assumed the energy from

joule heating is converted into H3 emission. to the sulfur and oxygen ion fluxes required by Wet al. (1997).

5.3. Particle Precipitation The scenario described above would appear to be ctent with the analysis of high resolution spectra of JupThe other most likely source of energy to produce thenonauroral regions obtained by Feldman et al. (1993)observed H3 emission is particle precipitation. This mecha-the Hopkins Ultraviolet Telescope (HUT). Their

nism has the advantage that it also identifies the source of concluded that, of the slit-averaged emission ratethe ionisation necessary to produce the ion. PrecipitationLyman- of 2.3 kR cm2, only about 1722% couof 1 to 10 keV electrons, with fluxes of around 1 to 10 ergsexplained by solar induced fluorescence. They furthes1 cm2, is usually invoked to explain the observed joviancluded that photoelectrons could not account for thauroral emission (e.g. Kim, Fox and Porter, 1992). Themainder of the emission and that other electron imelectrons are considered to originate mainly (i) from fieldmechanisms were required to supply the energy reqaligned currents which follow magnetic field lines near theto explain the observations. Further analysis by Morboundary between closed and open field lines, and (ii)et al. (1995) broadly supported Feldman et al.s (from the pitch-angle scattering along magnetic field linesinterpretation. The conclusion that electron impact/pof particles residing in Jupiters equatorial current sheet,itation is required to account for the observed UV emwith additional contributions, including some originatinghas recently been challenged by Liu and Dalgarno (1in the Io Plasma Torus.

however. These workers carried out detailed modellWaite et al. (1997) have recently proposed that the X-raythe HUT spectra and conclude that solar fluoresemission from Jupiters near-equatorial latitudes maycould account for 57% of the observed emission, withresult from energetic ion precipitation; their estimated re-toelectrons supplying the rest of the energy withoquirement is 0.8 ergs s1 cm2 of sulfur and oxygen ionscourse to any other source of excitation.with energies above 300 keV/amu. Waite et al. (1997) also

If Liu and Dalgarno (1996) are correct, then theresuggest that these ions can be supplied by precipitationsupport for particle precipitation to be found in the from the inner radiation belt, which might also supply thespectra. But their analysis may not be inconsistentkeV electrons deemed necessary for H3 production. Addi-the mid-to-low-latitude H3 emission being due in ptionally, results from the energetic particle instrumentparticle precipitation. Liu and Dalgarno derive a tem(EPI) aboard the Galileo atmospheric probe indicate thatture of 530 K, placing the emission peak for H2 molethere are high concentrations of energetic electrons ex-

above or within the homopause, normally placed artending all the way from the Io Plasma Torus to just 1.35 the 1 bar level. This conclusion is consistent withRJ, where their concentration appears to fall off abruptlyrequirement for moderate absorption of the emissi(Fischer et al., 1996), marking the inner boundary of themethane. But their analysis does not rule out precipiinner radiation belt in the jovian equatorial plane. Fischerparticles penetrating to the 1 to 10 nanobar level wet al. (1996) report that their electron flux profile is consis-the peak H3 production by solar EUV is predicted to tent with the measurements of Pioneer 11. The EPI results(Kim et al., 1992). In the auroral regions, precipitatingshow that MeV electron fluxes are around 106 cm2 sr1

electrons deposit most of their energy between 1 bas1 from 5 RJ to around 1.5 RJ, with a peak of nearly 107

100 nbar, just around the homopause, resulting in incm2 sr1 s1 around 2.2 RJ. Even though most of theH3 and atomic and molecular hydrogen emission. At particles detected by the EPI are trapped in radiation belts,latitudes, however, the dip angles of magnetic fieldthe Galileo results still suggest that there may be suppliesare more shallow than in the auroral regions and parof keV electrons available to leak into the jovian atmo-

following these field lines would indeed be expectsphere along field lines which correspond to the mid-to-deposit their energy higher in the atmosphere than clow latitudes. The observed H3 emission would require asponding auroral electrons do, leading to H3 emitotal flux between 0.1 and 1.0 erg s

1 cm2, dependingon the efficiency of ionisation and excitation. This flux is albeit with a lower intensity than in the aurorae. More

FIG. 4. Spatial distribution of the mid-to-low-latitude jovian H3 emission, E(H3 ), in erg cm

2 s1. (Corrections for line of sight effectbeen made.)

FIG. 5. Figure 4 overlaid with (a) footprints of the magnetic l-shells predicted by the Offset Tilted Dipole model of Acuna and Ness (b) contours of equal magnetic field strength; and (c) contours of constant dip angle. The dip equator is indicated by the line of dots. ((b) afrom Connerney, 1993).


10/11

66 MILLER ET AL.

entirely solar EUV produced H3 would produce a flat the temperature and E(H3 ) map, around III 170

coincides with a region where westward-drifting (B-glatitudinal emission profile (prior to line-of-sight correc-tion), to a first approximation, in contradiction to the ob- ent curvature) electrons would notice a steeply decre

magnetic field, lowering their mirror points and mserved concave profile.precipitation into the atmosphere more likely (the shield wiper effect). Conversely in the region bet6. COMPARISON OF E(H3 ) WITH MAGNETICIII 50 and 150, where westward-drifting elecFIELD PARAMETERSwould encounter a rapidly increasing field, particulathe northern hemisphere, H3 emission is at a reIf the bulk of the ionisation/excitation energy implied

by the observed H3 emission were being provided by parti- minimum.There appears to be some additional correlatiocle precipitation, one might expect the spatial variation of

E(H3 ) to be sensitive to magnetic field parameters. In Fig. tween the emission and the dip angle, shown in Fig. particular, the minimum in the emission follows th5a, the H3 emission parameter is shown overlaid with con-

tours of the parameter which distinguished various shells equator quite closely, particular for III 330 toThis result might be explained considering that the of magnetic field lines according to the maximum distance

they extend from Jupiters magnetic centre according to to which precipitating particles of any given energpenetrate will clearly depend on the angle at whichthe OffsetTilted Dipole model of Acuna and Ness (1976).

We shall call these l-shells, to distinguish them from the encounter the jovian atmosphere. Particles precipiat steeper angles would be more likely to penetramore usual L-shells which refer to distance from the gravi-

tational centre of the planet. In making comparisons be- deeper layers where concentrations of molecular hgen, and thus H3 production rates, are higher.tween the loci of the l-shells and the emission, it should

be noted (i) that the values of l refer to the maximumdistance of the field line from the location of the offset

7. CONCLUSIONStilted dipole, and (ii) that the l-shells are projected ontothe 0.6-bar planetary surface, whereas the ionospheric re-

The main conclusion of this study of the mid-togion of interest in this paper is about 1,000 km above this.

latitude emission of H3 is that this ion is responsibAs a result, (i) around III 180 the field lines of any l- a much greater cooling effect than previous modelinshell cross the magnetic equator about 0.1 RJ further from allowed for. We also conclude that these relativelythe gravitational center of the planet than the l number

values of E(H3 ) cannot be explained unless H3 is suggests, whereas around III 0 it is about 0.1 RJ less; transported from the auroral regions and/or there is

and (ii) all of the footprints at all longitudes should actuallycle precipitation in the mid-to-low-latitude regions, abe moved to slightly higher latitudes. One point of notebe necessary for the observed X-ray emission (Waite

is that the footprint of l 1.35 RJ only reaches down just 1997). The latter explanation, if correct, would suggesbelow latitudes of20 at some longitudes, while for others

particles with energies in the keV domain populatit is as high as 40. Most of the brightest emission occurs

space between the Galileo EPI cutoff at 1.35 RJ anat latitudes higher than the l 1.35 RJ footprint, as far as planet itself. Future models of the jovian ionosphere sour spatial resolution allows this to be determined, sug-

therefore take account of these effects. This and sgesting some correlation with the particle densities ob-

quent work (Miller et al., work in progress) suggestserved by the Galileo EPI. But levels of H3 emission for the mid-to-low latitudes temporal fluctuations onaround 0.1 ergs s1 cm2 still exist all the way to the jovian

scales of a few days are relatively minor. In the fuequator, suggesting that populations of keV particles may

therefore, we feel that spectroscopic studies of joviaexist inside the 1.35 RJ EPI cut-off for higher energy MeV emission such as the one presented here will be of electrons and positive ions. assistance in models of the non-auroral ionosphere

In Fig. 5b, E(H3 ) is shown overlaid with the magnetic in investigating links to magnetic field parameters anfield strength predicted by Connerneys O6 model (Con-

culatory phenomena.nerney, 1993). For any given latitude, there is a tendencyfor higher emissions to be associated with lower fieldstrengths, as would be expected from diffuse particle pre- ACKNOWLEDGMENTScipitation. Similarly, if one traces footprints of constant l

It is a pleasure to acknowledge the expert assistance of the (from Fig. 5a) in the northern and southern hemispheres,UKIRT, which is operated by the Joint Astronomy Centre, Hfor any longitude, there is also a tendency for H3 emissionbehalf of the U.K. Particle Physics and Astronomy Research C

to be higher in the hemisphere where the footprint of the (PPARC). PPARC is also thanked for the award of a Ph.D. studel-shell in question encounters the lower field strength. Even to HAL, and for Grant GR/K97837, which funded much of this

This paper has benefitted considerably from the comments of tmore noticeable, however, is that the warm channel of


11/11


referees: Dr. Sang J. Kim brought our attention to potential contaminants distribution of H3 in the jovian auroral ionosphere. J.Geophy97, 60936101.in the jovian reflected solar spectrum; an anonymous referee helped

clarify our comparison of H3 emission with magnetic field parameters. Kim, S. J., G. S. Orton, C. Dumas, and Y. H. Kim 1996. Infrared spSpecial thanks are also due to Prof. Renee Prange for her advice and copy of Jupiters atmosphere after the A and E impacts of Cassistance during the preparation of this paper and Paper II. We also ShoemakerLevy 9. Icarus 120, 326331.express our appreciation to Drs. Daniel Rego, J. Hunter Waite, Jr., and

Lam, H. A. 1995. Monitoring the jovian ionosphere using H3 emBrian Mitchell for helpful discussion during the preparation of this manu-

as a probe. Ph.D. thesis, University of London.script.

Lam, H. A., N. Achilleos, S. Miller, J. Tennyson, L. M. TraftonNote added in proof. It has been drawn to our attention that the Geballe, and G. E. Ballester 1997. A baseline spectroscopic st

vertical scale in Figs. 8a and 8b of Paper II should be divided by 3. the infrared aurorae of Jupiter. Icarus 127, 379393.

Lean, J. 1987. Solar ultraviolet irradiancevariations:A review,J. GeRes. 92, 839868.REFERENCES

Liu, W., and A. Dalgarno 1996. The ultraviolet spectrum of thedayglow. Astrophys. J. 462, 502518.Achilleos, N., S. Miller, J. Tennyson, A. Aylward, and D. Rees, 1997.

JIM: Time-dependent, three-dimensional modelling of Jupiters ther- Livingstone, W., and L. Wallace 1994. An Atlas of the Solar Spectmosphere. J. Geophys. Res., submitted. the Infrared from 1850 to 9000 cm1 (1.1 to 5.4 m). Nationa

Observatory Technical Report.Acuna, M. H., and N. F. Ness 1976. The main magnetic field of Jupiter.J. Geophys. Res. 81, 29172922. McConnell, J. C., and T. Majeed 1991. H 3 in the jovian ionos

J. Geophys. Res. 92, 85708578.Atreya, S. K. 1986. Atmospheres and ionospheres of the outer planetsand their satellites, pp. 139143. Springer-Verlag, Heidelberg. Miller, S., T. R. Geballe, L. M. Trafton, G. E. Ballester, and J. Te

1992. H3 distribution on Jupiter. Bull. Am. Astron. Soc. 24, 1034Atreya, S. K., and T. M. Donahue 1976. Model ionospheres of Jupiter. In

Jupiter(T. Gehrels, Ed.), pp. 304310. Univ. of Arizona Press, Tucson. Miller, S., H. A. Lam, and J. Tennyson 1994. What astronomy has lfrom H3 . Can. J. Phys. 72, 760771.Ballester, G. E., S. Miller, J. Tennyson, T. R. Geballe, and L. M. Trafton

1994. Latitudinal temperature variations of jovian H3 . Icarus 107, Miller, S., and 12 other authors, 1995. The effect of the impact of 189194. ShoemakerLevy 9 on Jupiters aurorae. Geophys. Res. Lett. 22

1623.Ben Jaffel, L., J. T. Clarke, R. Prange, G. R. Gladstone, and A. Vidal-Madjar 1992. A new model for the Lyman alpha bulge of Jupiter. Mitchell, J. B. A. 1990. The dissociative recombination of moleculGeophys. Res. Lett. 20, 747750. Phys. Rep. 186, 215248.

De Bergh, C., A. Martin., T. Owen, D. Gautier, J.-P. Maillard, B. L. Morrissey, P. F., P. D. Feldman, M. A. McGrath, B. C. WolveLutz, and P. Drossart 1992. Paper presented at workshop on variable H. W. Moos 1995. The ultraviolet reflectivity of Jupiter at 3.5A phenomena in jovian planetary systems, Annapolis, MD. tion from Astro-1 and Astro-2. Astrophys. J. 454, L65L68.

Clarke, J. T., and 19 co-authors 1995. HST far-ultraviolet imaging of Prange, R., D. Rego, L. Pallier, L. Ben Jaffel, C. Emerich, J. T. CJupiter during the impacts of Comet ShoemakerLevy-9. Science G. E. Ballester, and J. Ajello 1997. Detection of self-reversed L267, 13021307. lines from the jovian aurorae with the Hubble Space Tele

Astrophys J., in press.Connerney, J. E. P. 1993. Magnetic fields of the outer planets. J. Geophys.Res. 98, 18,65918,679. Rego, D., R. Prange, and J. C. Gerard, 1994. Auroral Lyman-

bands from the giant planets. 1. Excitation by proton precipitaDinelli, B. M., S. Miller, N. Achilleos, H. A. Lam, M. Cahill, J. Tennyson,the jovian atmosphere, J. Geophys. Res. Planets 99, 17,07517M.-F. Jagod, T. Oka, J.-C. Helico, and T. R. Geballe 1997. UKIRT

observations of the impact and consequences of Comet Shoemaker Roberge, W., and A. Dalgarno 1982. The formation and destrucLevy 9 on Jupiter. Icarus 126, 107125. HeH in astrophysical plasmas. Astrophys. J. 255, 489496.

Drossart, P., J-P. Maillard, J. Caldwell, S. J. Kim, J. K. G. Watson, W. A. Schultz, R., Th. Encrenaz, J. Stuwe, and G. Wiederman 1995. MonMajewski, J. Tennyson, S. Miller, S. K. Atreya, J. T. Clarke, J. H. the near-i.r. emission features at the NTT and detection of the noWaite, Jr., and R. Wagener 1989. Detection of H3 on Jupiter. Nature counterparts. In Proc. Euro. SL9/Jupiter Workshop, 1995 (R340, 539541. and H. Bohnhardt, Eds.), pp. 363368.

Feldman, P. D., M. A. McGrath, H. W. Moos, S. T. Durrance, D. F. Sommeria, J., L. Ben Jaffel, and R. Prange 1995. On the existeStrobel, and A. F. Davidsen 1993. The spectrum of the jovian dayglow supersonic jets in the upper atmosphere of Jupiter. Icarus 119observed at 3A resolution with the Hopkins Ultraviolet Telescope. Strobel, D., and S. Atreya 1983. Ionosphere. In Physics of the

Astrophys. J. 406, 279284. Magnetosphere (A. Dessler, Ed.), p. 57. Cambridge Univ. PressFischer, H. M., E. Pehlke, G. Wibberbez, L. J. Lanzerotti, and J. D. bridge, UK.

Mihalov 1996. High-energy charged particles in the innermost jovian Waite, J. H., Jr., T. E. Cravens, J. U. Kozyra, A. F. Nagy, S. K. Amagnetosphere. Science 272, 856858. and R. H. Chen 1983. Electron precipitation and related aeron

Hamilton, D. C., G. Gloeker, S. M. Krimigis, C. O. Bostrom, T. P. the jovian thermosphere and ionosphere. J. Geophys. Res. 88,Armstrong, W. I. Axford, C. Y. Fan, L. J. Lanzerotti, and D. M. 6163.Hunten 1980. Detection of energetic hydrogen molecules in Jupiters Waite, J. H., Jr., G. R. Gladstone, P. Drossart, T. E. Cravens, A. Mamagnetosphere by Voyager 2: Evidence for an ionic plasma source. W. S. Lewis, B. Mauk, and S. Miller 1997. Jovian equatorialGeophys. Res. Lett. 7, 813816. emission: Key to understanding high atmosphere temperature

sured by the Galileo probe. Science 276, 104108.Kim, Y. H., J. L. Fox, and H. S. Porter 1992. Densities and vibrational

Steven Miller et al- Mid-to-Low Latitude H3^+ Emission from Jupiter

Documents

Transcript of Steven Miller et al- Mid-to-Low Latitude H3^+ Emission from Jupiter