Io's oxygen source: Determination from ground-based observations and implications for the plasma torus

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 105, NO. A4, PAGES , APRIL 1, 2000 Io's oxygen source: Determination from ground-based observations and implications for the plasma torus W. H. Smyth Atmospheric and Environmental Research, Inc., Cambridge, Massachusetts M. L. Marconi Fresh Pond Research Institute, Cambridge, Massachusetts Abstract. From modeling analysis of two different ground-based observations for Io's neutral oxygen cloud in [O I] 6300 emission, new exobase atomic oxygen source rates produced by the composite effects of ion sputtering (represented by an incomplete collisional cascade distribution) have been determined ß to be 1.27 x 1028 and 1.47 x 1028 atom s -, ---4 times larger than earlier estimates. Implications for local heating and escaping energy rates of neutrals at Io and for ion pickup energy-input rates to the plasma torus produced by this larger oxygen cascade source and also by other atomic oxygen and sulfur source processes (approximately scaled from the oxygen cascade source) are discussed. The estimated power to sustain these O and S neutral sources is 0.84 x 10 2 W, including 0.88 x 10 TM for nonescape atmospheric heating. The total ion pickup input power to the plasma torus is estimated to be in the range x 10 2 W with an "outer source region" (neutral clouds above Io's exobase) to "inner source region" (charge exchange below Io's exobase) torus power input ratio of---1:1-2. A thin "power ribbon" created by the "inner source region" may provide a rationale for the observed ultraviolet "flashes and ø sparkles" of the S ++ plasma torus ribbon and the very abrupt variations of [O I] 6300 A emission observed at Io. The physical picture presented provides a viable solution to the mass loading and energy crisis of the plasma torus with the future determination of refined source rates important in establishing its validity. 1. Introduction leighs. In addition, in 1990, similar rocket measurements of the neutral sulfur cloud were also acquired in a large aperture Atomic oxygen from Io was first discovered in 1980 from using the Hopkins Ultraviolet Telescope [Durrance et al., ground-based observations by Brown [1981]. In this discovery 1995]. More recently, in 1993, Thomas [1996] acquired from measurement, very dim 8 +_ 4 Rayleigh [O I] 6300 emission ground-based observations at eastern elongation the only was detected from Io's extended neutral clouds in a small slit other [O I] 6300 emission measurement that has been publocated near (or at) the west ansa of the satellite's orbit and lished to date for atomic oxygen in Io's neutral cloud and reported well removed from the orbital location of Io. This emission, a similar brightness to that obtained earlier by Brown [1981]. which is thought to be excited by impact of plasma torus elec- In addition to these very dim observations of atomic oxygen trons, immediately indicated that a significant amount of and sulfur in the neutral cloud remote from Io, an increasing atomic oxygen had escaped the gravitational grasp of the satnumber of observations have more recently been made for ellite (contained within Io's Lagrange radius of satellite atomic oxygen and sulfur very near Io. These observations radii (Rio)) and had populated circumplanetary orbits. This have been almost entirely restricted to distances from Io that discovery indicated that atomic oxygen was also a constituent are within the Lagrange sphere of the satellite and hence of Io's local atmosphere and corona and that the ionization of sample oxygen in Io's local atmosphere and bound corona. these oxygen atoms would likely provide a significant O + Until recently, these near-io observations provided only aperplasma source for the heavy ion (O +, O ++, S +, S ++, S +++) Io ture-averaged brightness values that were much brighter than plasma torus that had been documented in 1979 by the Voythe dim emissions observed in the more distant neutral cloud. ager spacecraft mission. Rocket-based measurements in 1981 These aperture-averaged observations have included ultravio- [Durrance et al., 1983; Skinner and Durrance, 1986] of ultravilet emissions for atomic oxygen and sulfur obtained from the olet emissions from the plasma torus region using a large 15 x 110 arc sec aperture that was well removed from Io discovered International Ultraviolet Explorer (IUE) in [Ballatomic sulfur in Io's neutral cloud in 1425 emission with an ester et al., 1987; Ballester, 1989; M. A. McGrath, personal average brightness of 2.8 _+ 1.3 Rayleighs. These measurecommunication, 1995] and from the Hubble Space Telescope ments also confirmed the existence of atomic oxygen in the (HST) in [Clarke et al., 1994; Ballester et al., 1994; neutral cloud from a blend of 1304 emissions from atomic J. T. Clarke, personal communication, 1995; M. A. McGrath, oxygen and sulfur with an average brightness of 4.2 _+ 2 Ray- personal communication, 1995]. In addition, near-io observa- Copyright 2000 by the American Geophysical Union. Paper number 1999JA /00/1999JA tions of atomic oxygen in [O I] 6300 emission in an Io- centered 5.2 x 5.2 arc sec aperture have been obtained by an ongoing ground-based program [Scherb and Smyth, 1993; Scherb et al., 1999], which has now assembled a large systematic

2 7784 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS data set over the period from 1990 to In 1997, spectacular ultraviolet images of Io in a number of emission lines of atomic oxygen and sulfur were acquired with the new Space Telescope Imaging Spectrometer (STIS) aboard HST and reported by Roesler et al. [1999] with more of these STIS images for Io obtained in 1998 [Wolven et al., 1999]. In addition, two spectacular images of Io in atomic oxygen in [O I] 6300 emission were acquired from HST in 1997 when Io was eclipsed by Jupiter's shadow [Trauger et al., 1997]. For Io in eclipse, the disk-average brightness for atomic oxygen in opti- cal emission (6300, 6363 ) has also been measured in 1998 from ground-based observations (M. E. Brown, personal communication, 1998), and very high resolution color eclipse images were acquired in 1998 by the solid-state imaging (SSI) instrument aboard the Galileo spacecraft for which optical emission from atomic oxygen is the most likely candidate spe- cies for the observed brightness in the red ( ) filter [Geissler et al., 1999]. Only limited progress has been made in extracting the source rate for oxygen and sulfur at Io's exobase from the observed oxygen emission brightnesses noted above. The relationships between the neutral source rate and observed emission brightness involve a number of physical processes that are not fully understood. For Io's corona and neutral clouds these processes include the nature of the oxygen and sulfur source mechanisms at Io's exobase, the space-time nature of the neutral lifetimes in the near-io environment and plasma torus, and the nature of the excitation mechanism for atomic oxygen and sulfur and their spacetime dependence in the near-io environment and plasma torus. For emission from the local atmosphere the situation becomes even more complex and involves factors that include the location and activity of volcanic gas plumes, the dynamics of the combined volcanic and sublima- tion atmospheric gases heated by plasma processes, additional,excitation produced by the Birkeland-current closure at Io, and atmospheri chemistry caused by solar photons, gas-phase collisions, and gas-plasma interactions. For the dim oxygen neutral cloud, progress has nevertheless been made by drawing physical insight from studies for the spatial nature of the sodium cloud determined from ground- based observations of the bright D lines (5890, 5896 ), which are at least 2 orders of magnitude brighter than optical emission for atomic oxygen (see review by Smyth [1992]). An atomic oxygen source rate of 3.2 x 1027 atoms s - was reported by Scherb and Smyth [1993] for the observation of Brown [1981] by assuming (1) an isotropic and 2.6 km s -l monoenergetic source at Io's exobase, which had proven useful to characterize approximately the morphology of the neutral sodium cloud near Io's orbit, and (2) an electron impact excitation mechanism in the plasma torus. Since that time a complete source velocity distribution has been determined for sodium at Io's exobase from modeling of combined sodium observations on several different spatial scales [Smyth and Combi, 1997]. The dominant source mechanism that populates the neutral cloud near Io's orbit was shown to be an incom- plete collisional cascade (sputtering) source distribution that peaks at 0.5 km s-, well below the satellitescape speed, and extends in the tail of the distribution to many tens of kilometers per second. With this new insight into the source velocity distribution for the neutral cloud for sodium, which is a trace species in Io's local atmosphere and corona, it is now possible to adopt this source velocity distribution as a first-order description for the more dominant but very dim oxygen and sulfur atoms located in the vicinity of Io and near its orbit and proceed to analyze the neutral cloud observations to determine more realistic source rates. The optical observations of Brown [1981] and Thomas [1996] for atomic oxygen are ideally suited for this analysis because the electron impact excitation rate is well known and the observing slit is relatively small and dominated by emission near Io's orbit populated by atoms described by the adopted source velocity distribution of Smyth and Combi [1997]. This analysis is the objective of this paper. An analysis of the equally interesting ultraviolet observations for Io's neutral O and S clouds [Durrance et al., 1983; Skinner and Durrance, 1986; Durrance et al., 1995] is, however, less straightforward and will not be considered here because of current uncertainties in atomic parameters determining the electron impact emission rates for some of these ultraviolet lines and because of the large field of view of the aperture, which would likely introduce the additional modeling uncertainties of describing emission contributions from the spatially huge O and S zenocoronae in the distant foreground and background. The paper is organized as follows. The observations of Brown [1981] and Thomas [1996] are presented in section 2. In section 3 the neutral cloud model used in the analysis is described followed by its application to the observations. A discussion of the model results, their implications, and conclusions is given in section Observations Ground-based observations reported by Brown [1981] and Thomas [1996] for Io's neutral oxygen cloud in the 6300 emission line were both acquired by a slit with a field of view that was well removed from Io's immediate location. The sky plane geometry for these two observations is presented in Figure 1 and shows the relative locations of the observing slits, Jupiter, Io, and Io's orbit. The observation of Thomas was acquired exactly 13 years to the day after the observation of Brown, which is close to an orbital period of Jupiter and for which the two tilt angles for Io's orbit projected onto the sky plane were similar. The observations of Brown [1981] were made on March 5, 1980, with a small slit that covered a 0.32 x 0.05 Rj rectangle with the long side parallel to the satellite orbital plane. As illustrated in Figure 1, the observing slit of Brown was centered on the symmetry line of the projected sky plane ellipse of Io's orbit and was positioned at 5.0 R west (right) of Jupiter for two observations and at 5.9 R west (western elongation) of Jupiter for four observations. During these six measurements the satellite was moving toward western elongation (270 ø ) on its orbit and covered an Io g½ocentric phase angle interval of 181ø-215 ø. The small solid circles in Figure 1 to the right of Jupiter show the orbital locations of Io at the midpoint time of each of the six observations. During this time, Io also covered a System III longitude interval in the plasma torus from 105 ø to 215 ø, so that the satellite quickly crossed the plasma torus centrifugal symmetry plane (110 ø) and then spent the remainder of its time north of that plane. The six observed brightnesses were properly Doppler-shifted and added by Brown to improve signal to noise and provided a slit average [O I] 6300 emission brightness of 8 +_ 4 Rayleighs. The primary [O I] 6300 emission observation (number 199) reported by Thomas [1996] was acquired on March 5, At the midpoint of this observation, Io was approaching eastern elongation (90 ø) with a geocentric phase angle of 73 ø

3 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS i 2 Thomas DISTANCE (R j) Figure 1. Geometry for the [O I] 63003, observations near Io's orbit. The March 5, 1980, observations of Brown [1981] were made with a 0.32 x 0.05 Rs slit (small rectangles with the long side parallel to the orbital plane) located for two observations 5.0 R s west (right) of Jupiter (circle at the center) and for four observations 5.9 R s west of Jupiter so as to be centered on the elongation point of the satellite's orbit. Io's orbit projected onto the sky plane is shown by the thin ellipse. The positions of Io at the midpoint of the six observations of Brown are shown by the small solid circles. The March 5, 1993, observation of Thomas [1996] was obtained with a vertically oriented slit (rectangle on the left side of Jupiter) of effective dimensions 7 x 0.07 Rj that was located close to the position of Io (indicated by the small open circle). (small open circle left of Jupiter in Figure 1) and was south of plane as shown in Figure 2 with a peak brightness of 8.8 _+ 1.7 the plasma torus centrifugal symmetry plane with a System III Rayleighs. This peak brightness is similar to the 8 _+ 4 Raylongitude of 339 ø. As illustrated in Figure 1, the observation leighs 63003, average emission brightness measured by Brown was acquired using a slit of dimensions 7 x 0.07 Rs with the [1981] for the forward oxygen cloud near the western elongalong axis oriented perpendicular to Io's orbital plane and tion point of Io's orbit. A second but poorer [O I] 63003, placed near the eastern elongation point of Io's orbit and emission observation (number 212) acquired 2.5 hours later Jupiter radius (---20 Io radii) from the center of the disk of Io. was also reported by Thomas for Io with a midpoint Io geo- The slit is so located to measure Io's forward oxygen cloud. A centric phase angle of 95 ø and an Io System III longitude of 51 ø reliable north-south brightness profile for the 6300 emission (i.e., the satellite was still south of the plasma torus equator line along the slit was obtained only very close to the satellite symmetry plane). For this measurement a similar peak emission brightness of 9 _+ 3 Rayleighs was obtained but was located 0.19 Rs ( Io radii) north of Io for a north-south oriented slit centered on the satellite. In this paper, modeling analysis of the Thomas [O I] 6300 emission observations will be limited to the better and more fully described observation (number 199) shown in Figure I... I DISTANCE ALONG THE SKYPLANE SLIT (R j) 3. Analysis of the Observations 3.1. Model Description The mathematical formulation and description of the neutral cloud model for atomic oxygen adopted here were developed by Smyth and Combi [1988a], and the application of this model to the oxygen observations is similar in style and in treatment of the plasma torus to that adopted in the earlier application to the sodium neutral cloud [Smyth and Combi, 1988b]. In short, atoms are ejected from the satellite exobase with velocities that describe a given source velocity distribu- tion, and their trajectories in space are determined by solving Figure 2. [O I] 6300 observational profile of Thomas the circular restricted three-body equations of motion that [1996] and comparison with model calculation. The observed north-south spatial profile for the [O I] 6300 brightness include the gravity fields of Io and Jupiter. Along these trajecdistribution is shown by the asterisks with vertical-line error tories, appropriate weights are calculated to determine in three bars included. The negative numbers are south of the orbital dimensions the space- and time-dependent gas density, voluplane. The model-calculated brightness profile is shown by the metric emission rates, and ion production rates caused by the solid line. interaction of the gas and the oscillatory motion of the plasma

4 7786 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS Table 1. O Loss Processes in the Plasma Torus Near Io Relative Rate, Lifetime, % hours O + e -- O + + 2e O + O +- O + + O O + O ++ - O ++ + O > , S + --> S ,000 O + S ++ - O + + S O + S (O +)* + (S + +)* O + h v -- O + + e ,000 Total with (1) ballistic trajectories (-62%) that populate only the corona of the satellite, (2) low-speed escape trajectories that have similar speeds to the monoenergetic 2.6 km s - source adopted earlier by Scherb and Smyth [1993], and (3) higherspeed escape trajectories that quickly move away from Io's orbit and do not contribute significantly to the forward cloud. For sodium the source speed distribution is responsible for the creation of the very steep gradient near Io in the spatial distribution of sodium within Io's corona as well as the more gradual gradients in the forward sodium cloud and the trailing sodium cloud (excluding the higher-speed contributions by the directional feature source) near Io's orbit. Higher-velocity neutral source processes (i.e., direct ejection and charge exchange) that are responsible for creation of the sodium directional torus relative to Io's orbital plane. The three-dimensional de- feature and the sodium zenocorona are, as would be expected, scription of the electron and ion densities and temperatures in anisotropic with a preferential direction aligned with the the plasma torus, adopted in the neutral cloud model, is based plasma corotational motion. By adopting this more realistic on a tilted and offset dipole planetary magnetic field, the incomplete collisional cascade source distribution at the exoplasma properties near western elongation for the 1979 Voyager base and requiring the same number of atoms to populate the spacecraft encounter, a nominal east-west electric field, and a forward cloud near Io's orbit, the oxygen source rate at the System III longitude asymmetry in the active sector. Because exobase will, necessarily, be larger than the value determined of the east-west electric field, the plasma torus properties are earlier by Scherb and Smyth [1993] for the observations of also time varying in local time [see Smyth and Combi, 1988b]. Brown [1981] when using only a simple 2.6 km s - monoener- The source velocity distribution adopted at the exobase for getic source. atomic oxygen is the isotropic incomplete collisional cascade The lifetime of oxygen atoms is determined by electron distribution determined for sodium by Smyth and Combi [1997] impact and charge exchange processes in the plasma torus. which represents the composite effects of ion sputtering, in- Both processes are inherently space-time dependent, and the cluding heating of the atmosphere. The distribution is denoted charge exchange rate is also velocity dependent because of the "incomplete" because the asymptotic tail power-law behavior relative motion of the inhomogeneous plasma torus and the of E -4/3, where E is the kinetic energy of a neutral, is over- atom trajectories. The charge exchange cross sections adopted populated relative to a complete collisional cascade distribu- in the neutral oxygen cloud model are those given by McGrath tion with an asymptotic power-law behavior of E- 2. This over- and Johnson [1989] with refinements provided by R. E. Johnpopulation occurs because of the lack of the completion in Io's son (personal communication, 1989). The relative importance atmosphere (where escape occurs at the top) of the collision of the different lifetime processes at Io's orbit in the plasma cascade process that moves energy from the velocity tail to the torus symmetry plane at western elongation is summarized in lower-velocity core of the distribution. The incomplete colli- Table 1. At this location, charge exchange is seen to dominate sional cascade flux distribution 4>(v; a, vt,, vm), expressed in over electron impact ionization. The relative importance of the terms of the neutral speed v, is discussed in detail by Smyth lifetime processes in Table 1 will, however, vary for different and Combi [1988b, Appendix D] and corresponds to an asymp- locations in the plasma torus. The minimum charge exchange totic tail power-law behavior of E - ("- ) and has a peak (most lifetime, for example, is located near Io's orbit where the ion probable) speed of Vm that is related nonlinearly to the velocity densities are largest. The minimum electron impact lifetime, parameters vt, and va4. The core of the source distribution at however, occurs near 7 Jupiter radii because of the combined v = Vm provides an effective parameter to characterize at Io's effects, with larger radial distances, of the increasing electron exobase the composite effects of the complex atmospheric temperature and the decreasing electron density. The total lifeheating processes that occur below the exobase. For the in- time of atomic oxygen in the plasma torus coordinate frame is complete collisional cascade process, this source distribution shown in Figure 3 in a two-dimensional plane at eastern elonqhcn ld alqc anniv to atomic oxygen, at lea,qto first order. since gation for the time of the Thomas [1996] observation and at sodium is a trace species. The distribution was determined western elongation for the time of the Brown [1981] observation. from a set of sodium observations to be isotropic, with a peak The [O I] 63003, emission for oxygen atoms along the atom at Vm = 0.5 km s- (well below the exobasescape speed of trajectories is assumed to be excited by plasma torus electron -2 km s -q) and a tail characterized by a = 7/3 that extends to impact and is hence also inherently space-time dependent. The many tens of kilometers per second (see Smyth and Combi electron impact [O I] 6300 volumetric emission [1997] for the high sensitivity of the fit). For these parameter in the neutral cloud model is well established, and its depenvalues the average neutral energy of the flux distribution as dence upon the electron temperature is shown in Figure 4. applied to atomic oxygen and sulfur was recently determined This emission rate is calculated using the Collisional and Raby Smyth [1998]. About 62% of the source atoms for this diative Equilibrium (COREQ) code with updated atomic data distribution are on nonescape trajectories, while the remaining files developed by D. E. Shemansky (personal communication, -32% escape Io's gravitational grasp. The composite pro- 1992). The local emission photon rate per oxygen atom is cesses below the exobase that produce this incomplete cascade obtained by multiplying this emission rate by the local electron distribution are yet to be understood from first principles by density. The line of sight integration of the product of the the solution to a complex problem involving the collisional neutral oxygen density with the electron density and the emisatmosphere and its interaction with the plasma torus. The sion rate then determines the neutral cloud brightness on the source distribution thus provides a significant source of atoms sky plane of the observer. rate adopted

5 ._ SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS ',-i...!... i... i... i... i.... ' "" F...,...,...,......,...,...-I ß 0 o._ -3,,, i i i i I,..i.,, Plasmacentric Radial Distance (R j) Plasmacentric Radial Distance (R j) Figure 3. Lifetime of atomic oxygen in the plasma torus. The total (electron impact plus charge exchange) lifetime in units of hours for atomic oxygen is shown in the coordinate frame of the plasma torus in (a) at eastern elongation for the observation of Thomas [1996] that corresponds to an east ansa System III longitude of 316 ø and in (b) at western elongation for the midpoint of Brown's [1981] observation that corresponds to a west ansa System III longitude of 88 ø. The plasmacentricoordinate frame, used earlier by Smyth and Combi [1988b], includes the ---7 ø tilt of the plasma torus centrifugal equator with respecto the orbital plane of Io, a dipole offset of the magnetic field relative to the center of Jupiter, and the presence of an east-west electric field of mv m - as seen in the Io frame. At the two elongation points the locus of Io's position in the plasma torus coordinate frame is shown by the light gray oval-like closed line on which the oxygen lifetime varies periodically as a function of the System III magnetic longitude of the satellite Modeling Analysis Model calculations for the two-dimensional [O I] 6300, emission brightness on the sky plane have been undertaken using the atomic oxygen neutral cloud model described above with the more realistic incomplete collisional cascade source distribution. The source rate at the satellite exobase for atomic oxygen is determined by 'matching the model-calculated and 10-9 z o measured brightness in the observing slit. For the March 5, 1980, observation of Brown [1981], the model calculation is shown in Figure 5 that corresponds to the midpoint location of Io for the six co-added observations. The model calculation matches the slit average [O I] 6300, emission brightness of 8 _+ 4 Rayleighs for an isotropic oxygen exobase source of 1.27 x 1028 atoms s -. For the March 5, 1993, observation of Thomas [1996], the oxygen model calculation is shown in Figure 6 for the midpoint of the observation. An isotropic exobase oxygen source of 1.47 x 1028 atom s- is required to match the average brightness of 5.7 Rayleighs in that portion of the observing slit in Figure 2 for which a measured signal was acquired. The model-calculated and measured north-south brightness profiles for the Thomas observation are shown in Figure 2 and compare reasonably within experimental error. There is an apparent discrepancy in the location of the peak intensity between the model and data. Given the substantial experimental error, however, we will leave an analysis of this for a future study of more recently obtained, but not yet avail- able, data for atomic oxygen acquired by N. Thomas (personal communication, 1998) ,,,,,, 1 I I ELECTRON TEMPERATURE (K) Figure 4. Electron impact emission rate for the [O I] 6300, line. The electron impact emission rate for the [O I] 6300, line as a function of electron temperature is shown as determined from the Collisional and Radiative Equilibrum (COREQ) code developed by D. E. Shemansky (personal communication, 1992) and based on atomic physical parameters from Wiese et al. [1966], Henry and Burke [1969], and Moore [1971]. 4. Discussion and Conclusions 4.1. Summary Discussion for the New Oxygen Rates The atomic oxygen source rates of 1.27 x 1028 and 1.47 x 1028 atoms s - determined from the two observations are sur- prisingly very similar despite the 13 year time interval between these measurements. Since only about 0.38 of the oxygen atoms escape Io for the incomplete collisional cascade source distribution adopted here, the source rates for the escaping fraction are 4.83 x 1027 and 5.59 x 1027 atoms s -1, respectively, for the Brown [1981] and Thomas [1996] measurements. The estimated escape rate of Scherb and Smyth [1993] for an

6 7788 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS 3 2- I. I, I i I, I, I i ' DISTANCE Rj Figure 5. Model calculation for Brown's [19811 observation of the sky plane brightness of the [O I] 6300 emission. The contours of sky plane [O I] 6300 A brightness were modeled with a single calculation for Io at the midpoint (a satellite phase angle of 198 ø and an Io System III longitude of 161 ø) of Brown's six observations. Atomic oxygen was assumed to be ejected isotropically from Io's exobase (radius assumed to be at 2600 km) with an incomplete collisional cascade source characterized by a - 7/3 and a most probable speed of 0.5 km s -j, as determined by Smyth and Combi [1997] for atomic sodium at Io's exobase. The contour levels are, from outside to inside, 1.0, 2.0, 5.0, 10.0, and 20.0 Rayleighs. The total oxygen flux in the model was normalized so as to give the brightness of 8 Rayleighs reported by Brown. Also shown are the locations of Jupiter, Io's orbit, and Brown's observing slit (small rectangles). analysis of Brown's observation, assuming a simpler monoen x 102s atoms s - is adopted. This average rate provides ergetic 2.6 km s- initial velocity for the oxygen at Io's exobase, an absolute normalization base upon which cross-sectional was 3.2 x 1027 atom s-. Hence the new oxygen escape rates scaling arguments can be used to estimate roughly the incomare times larger, and the total new oxygen exobase source rate (escaping and nonescaping) is -4 times larger. The plete collisional cascade source rate for atomic sulfur and, in addition, the source rates for oxygen and sulfur for the pronew atomic oxygen escape rates are only -5-21% larger than cesses of low-velocity (-20 km s - ) charge exchange (and the escape rate of 4.6 x 1027 atoms s - for the incomplete direct ejection) and full-velocity (nominally-60 km s - ) collisional cascade estimated more recently by cross-sectional charge exchange. The scaling procedure, employed earlier by scaling arguments [Smyth, 1998]. It is interesting to note that &nyth [1992, 1998], provides a useful although approximate the new atomic oxygen total source rate is also similar to the tool to project the oxygen results forward into the more cominferred rate obtained by assuming two oxygen atoms per SO2 plete picture and to explore likely implications. For the three molecule for the SO2 source rate of 8.9 x 1027 molecules- collisional source processes noted above, the oxygen and sulfur estimated by Scherb and Smyth [1993] from near-io [O I] 6300 emission measurements. source rates obtained by the scaling procedure are summarized in Table 2. The procedure is basically to assign appropriate 4.2. Power Input Estimates for Neutrals and Ions torus-averaged cross sections to the three different collisional and Their Implications The new oxygen source rates for the incomplete collisional cascade distribution at Io's exobase have interesting implications. To explore these implications more concretely, the average oxygen exobase source rate for the two measurements of processes, assume the individual source rates are proportional to their cross sections, and then set the proportionality constant equal to the escape component (38%) of the incomplete collisional cascade source rate divided by the cross section for source atoms emerging in the neutral speed range 2-10 km -3. i i i i i, i i i i i l DISTANCE Rj igure 6. Model calculation for Thomas's [1996] observation of the sky plane brightness of the [O I] 6300 emission. Same as Figure 5 except that the model corresponds to the geometry of Thomas's observation number 199 with an Io phase angle of 73 ø and an Io System III longitude of 339 ø at the midpoint of the observation. The model flux was normalized so as to reproduce the average brightness of 5.7 Rayleighs in that portion of the slit included in Figure 2.

7 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS 7789 Table 2. Neutral Source Rate Scaling Procedure Collisional Source Process Neutral Speed Range for Cross Section, km s -l Atomic Oxygen Atomic Sulfur Cross Section," Source Rate, ' Cross Section, ' Source Rate, t' ]0- (' cm atoms S cm atoms s - Incomplete collisional cascade Nonescape (62%) 0-2 Escape (38%) 2-10 Low-velocity charge exchange/direct ejection >10 Full-velocity charge exchange -60 ß ß ' The torus density-averaged crossections are taken from Smyth and Combi [1991, Appendix A] with the nonthermal crossections for the speed range indicated assigned to the incomplete collisional cascade processes and to the low-velocity charge exchange/direct ejection collisional source processes. t'thescape component (38%) of the incomplete collisional cascade source is assigned to the torus-averaged crossection for the neutral speed range between 2 and 10 km s-, and the remaining source rates are determined from this rate by scaling with the assigned crossections. The incomplete collisional cascade source adopted for atomic oxygen is 1.37 x 1027 S- l, the new average value obtained this paper, with one half of this new oxygen source rate assumed for the incomplete collisional cascade source rate for atomic sulfur. s -1. In Table 2 the incomplete collisional cascade source rate 1.0 x l0 2 W [Smyth, 1992], respectively. The input power to for atomic oxygen of 1.37 x 1028 s -1 is adopted, and the sustain all neutral processes is estimated below to be 0.84 x corresponding rate of 6.85 x 1027 S -1 is assumed for atomic 1012 W if ion heating of Io's atmosphere is also included. For sulfur, which is one half of the new average oxygen rate, as the three source processes in Table 3, the individual rates and would be appropriate if SO2 were the parent molecule. their spatial nature are discussed separately below. The exobase source rates in Table 2 and their corresponding For incomplete collisional cascade, the total estimated exoexobasescape rates (38% for incomplete collisional cascade base source rate for O and S in Table 3 of ---2 x 1028 atoms s -1 and 100% for the other source processes) are summarized in is produced primarily by multiple collisional processes deeper Table 3 for the three collisional source processes and are used in Io's atmosphere and has essentially the same value as the to estimate ion pickup rates and two energy input rates. The total estimated exobas escape rate of O and S for all source "total ion pickup rate" (kg s -l) in Table 3 is determined by the processes. The power of 0.22 x l0 TM W required to sustain the exobasescape rates for oxygen and sulfur. The values for the incomplete collisional cascade distribution a significant frac- "torus energy input rate" (the net rate that gyrokinetic energy tion (33%) of the initial upstream ion corotational kinetic is supplied to new pickup ions in the plasma torus through energy rate (0.67 x l0 TM W) flowing through an Io disk area ionization and/or charge exchange of neutrals for each source but is small compared to the instantaneous new pickup ion net process) are determined from the "total ion pickup rate" by gyrokinetic energy rate of 4.99 x l0 TM W that these neutrals assuming simply a nominal pickup speed of 60 km s --, al- add to the plasma torus energy budget. It should be noted that though the power for the two higher-velocity processes, as the power extracted from the corotating plasma to sustain this discussed later, is likely larger. The "neutral energy input rate" neutral source is highly localized in Io's atmosphere. This is in (the rate that kinetic energy is supplied to neutrals by ions for contrasto the instantaneous power input to the plasma torus each collisional source process) is determined by multiplying (and total pickup rate of 277 kg s -1) provided by this neutral the exobase source rate for each atom species by the average source which is located, at the time of creation, in the spatially atom energy of the source distribution (which was determined distributed volume of the extended neutral clouds so as to for these three processes by Smyth [1998, Appendices A and produce an "outer source region" for the Iogenic plasma (i.e., B]) and summing the two species input rates for that source existing above and well beyond Io's exobase both ahead and process. In summary for all three processes in Table 3, the total behind Io and concentrated near Io's orbit but still highly ion pickup rate of 664 kg s -1, the neutral escape-energy input peaked about Io, as described by Smyth and Marconi [1999]). rate of 0.75 x 10 TM W, and the total torus energy input rate of The instantaneous power source introduced by these pickup at least 1.20 x 1012 W are a little larger than earlier values of ions subsequently, of course, will move relative to Io with the 592 kg s -1 [Smyth, 1992], 0.65 x 1012 W [Smyth, 1998] and plasma flow and provide a circumplanetary energy source for Table 3. Summary of Source and Energy Input Rates Collisional Source Process Incomplete collisional cascade Low-velocity charge exchange/direct ejection Full-velocity charge exchange Total Exobase Source Rate Oxygen, Sulfur, atoms s- 1 atoms s- Exobase Escape Rate Neutral Torus Total Ion Energy Energy Oxygen, Sulfur, Pickup Rate, ' Input Rate, Input Rate, t' atom s - atom s-- kg s- W W 1.37 x X x X x 10 TM 4.99 x 10 TM x x X X x 10 TM >1.73 x 10 TM x X X X x 10 TM _>5.24 x 10 TM 2.25 x x x x x l012 2>1.20 x l012 ' Based on the oxygen and sulfur exobase escape rates and thus the total mass loading rate for all electron impact and chargexchange processes. t'the Torus Energy Input Rate is obtained simply from the Total Ion Pickup Rate (kg s - ) by assuming a nominal pickup ion speed of 60 km s -

8 7790 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS the plasma torus that is radially broadly distributed about Io's orbit. The power of 0.22 x 10 TM W required to sustain the incomplete collisional cascade exobase source rate and hence this circumplanetary distributed energy source, however, has been estimated to be only ---20% of the power extracted locally from the plasma flow by Io's atmosphere through elastic ionmaining unexplained nt reduction in the local planetary magnetic field measured near Io by the Galileo spacecraft [Kivelson et al., 1996a, b; Khurana et al., 1997]. The charge exchange neutral source is accompanied by an instantaneous plasma torus energy input rate of ->5.24 x 10 TM W (for a pickup speed of ->60 km s - ) that is highly localized in Io's sputtering interaction processes [Pospieszalska and Johnson, "inner source region" and that, in effect, provides a very con- 1996; Marconi et al., 1996]. The remaining ---80% of the power centrated point energy source for the plasma torus. Upon (0.88 x 10 TM W), which is deposited as heat in Io's atmosphere, subsequently moving downstream of Io, this point source will is important in raising Io's exobase altitude and altering the day-to-night sublimation gas flow [Wong and Johnson, 1996] and is primarily lost by radiation and heat conduction/ form a very concentrated energy source thread that will rapidly change into a very concentrated energy source ribbon (in magnetic latitude and System III longitude). Prior to significant convection to the surface. The total estimated ion-sputtering radial transport, this pickup-ion "power ribbon" will have a power input to Io's atmosphere would then be x 10 TM W and is very similar to values obtained in recent twoheight ---2 Rj and a width of <0.05 Rj, which is ->4 times thinner than the nominal 0.2 R width of the S + plasma torus dimensional modeling of Io's atmospheres by Wong and Smyth ribbon [Schneider and Trauger, 1995]. [2000] and in Monte Carlo ion sputtering studies for Io's at- The power ribbon will wander initially a'nonuniform manmosphere by Lisiecki et al. [1999]. For low-velocity charge exchange and direct ejection, the ner within the L-shell volume of the plasma torus created by the pickup ion trajectories as determined by the combined combined O and S exobase source rate in Table 3 of 2.77 x effects of the corotational, east-west electric field, and System 1027 atoms s - is produced at those higher altitudes in Io's atmosphere (and ionosphere) where ions and (newly created III longitude-offset motions [Smyth and Marconi, 1998] and will subsequently undergo radial plasma transport within and or secondary collisionally produced) atoms with intermediate beyond this volume. The energy in the power ribbon is ultispeeds (---20 km s - ) can more easily travel and escape without mately transferred to the electrons and through electron-ion sustaining collisions. Although this exobase source rate is almost an order of magnitude smaller than the incomplete colinelastic collisions is lost by radiating ion emission lines. The cumulative degree of nonuniformity of the power ribbon and lisional cascade source, the neutral power input of 0.37 x 10 TM the radiated emissions within this volume will depend upon the W required to sustain this smaller source is larger because the energy per atom for this distribution is a factor of larger. Whether or not the low-velocity charge exchange and relative timescales for energy decay and radial transport. For somewhat comparable timescales, as may occur for collisional decay of the pickup ions with background ions and electrons direct collisional ejection source of 96 kg s- provides a small [Barbosa et al., 1983] having a lifetime of approximately a few net gain or loss for the gyration energy of the involved ions in this "inner source region" of Io's atmosphere (i.e., the spatial region below the exobase) is difficult to assess. Because of their larger neutral velocities, these escaping neutrals will, however, occupy in the "outer source region" a much larger volume of the inner magnetosphere than that of the incomplete collisional cascade source primarily localized near Io's orbit and provide an estimated instantaneous power input to the plasma torus budget of >1.73 x 10 TM W (since their pickup speeds at these larger distances will be higher than 60 km s-i). This larger spatially distributed power source for the plasma torus 106 S, a somewhat spatially uniform power source is expected. If, however, the dominant energy sink is controlled by shorter timescale processes, such as may occur for wave-particle interactions, the power ribbon may be expected to remain very concentrated and nonoverlapping so as to produce a highly nonuniform volumetric energy source. Such a highly nonuniform volumetric distribution for the pickup ion power ribbon would indeed provide a rationale for the short-term temporal variability or time-dependent "flashes and sparkles" of the ultraviolet 685 brightness of the torus S ++ plasma ribbon observed in the Voyager ultraviolet spectrometer (UVS) data may then be further enhanced by inward transport processes, [Sandel and Broadfoot, 1982; Volwerk et al., 1997], since the as described by Smith et al. [1988]. For full-velocity charge exchange, the combined O and S exobase source rate in Table 3 of 8.91 x 1027 atoms s - is produced at those upper altitudes in Io's atmosphere near and below the exobase where the fast corotational ions and newly created atoms can travel and escape without sustaining a secemission rate for these 18.1 ev ultraviolet photons is extremely sensitive to the high-energy tail of the electron distribution function. Likewise, a highly nonuniform distribution might also play a role in producing the very abrupt temporal variability of the [O I] 6300 emission brightness observed in a 5.2 x 5.2 arc sec aperture centered on Io for nearly a decade from ond collision. These neutrals rapidly escape the Jupiter system ground-based measurements [Scherb and Smyth, 1993; Scherb and are expected to produce extremely dim zenocoronae for O and S [Smyth and Combi, 1991]. Only the much brighter atomic sodium zenocorona has been observed to date [Mendillo et al., 1990]. Because of their large speed, however, they will provide almost no spatially extended plasma power source for the "outet al., 1999] but to a much lesser extent since the emission rate for these ---2 ev optical photons is much less sensitive to the electron temperature than the S ++ ultraviolet photons. Furthermore, such a highly nonuniform distribution for the power ribbon would in addition be consistent with the observed shorter source region" in the inner magnetosphere. The power of time stability of the optical S + electron impact emission lines 6.89 x 10 TM W required to sustain the chargexchange neutral (6716 and 6731 ) in the plasma torus, since source (and which simultaneously provides an ion pickup rate of 291 kg s - for the plasma torus) is so large that it can be supplied only through the local planetary magnetic field energy that is partially converted to neutral kinetic energy by the ion pickup current that it creates. As discussed earlier by Smyth [1998], this ion pickup current is the likely source of the relines are very insensitive to the electron temperature Implications for the Plasma Toms Energy Budget The energy budget for the plasma torus, which is dominated by the ultraviolet energy radiated by the heavy sulfur ions, is estimated to be approximately a few x10 TM W [Hall et al., 1995] thesemission

9 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS 7791 and is only about a factor of 2 or so larger than the estimated supply the additional energy input to the plasma torus that total torus energy input rate of at least 1.20 x 1012 W in Table could not be supplied by the "neutral cloud theory" as approx- 3. The contribution to this estimated total torus energy input rate by the incomplete collisional cascade source (42%) loimately described in a one-box model where both the nominal (and homogeneous) neutral gas densities and plasma torus cated in the "outer source region" is reasonably firm, being properties were simultaneously required to be obtained as a basedirectly upon the [O I] 6300 emission observations of steady state solution of the coupled continuity and energy Brown [1981] and Thomas [1996]. The contributions to the equations. In order to provide an appropriate increase in the estimated total torus energy input rate by the low-velocity electron temperature and a proper value for the S ++ to S + charge exchange and direct ejection source (14%) located in abundance ratio, Shemansky [1988] determined that a torus the "outer source region" and the full-velocity charge exchange power input ratio of -- 1:2 was generally needed for the neutral source (44%) located in the "inner source region" are, however, determined from cross-sectional scaling arguments. The cloud "outer source region" to charge exchange "inner source region" ratio. This power input ratio is indeed similar to the scaling arguments are only a first-order approximate means of recent MHD simulation ratio of -- 1:2-1:3. Hence a viable estimating these two higher-velocity source rates. The rates solution to the mass-loading and energy crisis [Shemansky, cannot be determined directly from the very dim [O I] ; Smith et al., 1988] of the plasma torus is that the plasma emission brightness for these two higher-velocity source rates since they are well below current observational detection limits. Alternatively, a direct theoretical calculation of the two higher-velocity source rates is well beyond current modeling capabilities. In addition to expected time dependence, the contorus mass is supplied by the extended neutral clouds and possibly some direct ion escape from Io and that the plasma torus energy is provided primarily by the pickup ion power ribbon created by localized charge exchange processes in Io's atmosphere near and below the satellite exobase. tribution to the torus energy input rate of the full-velocity charge exchange source for O and S alone in Table 3, however, 4.4. Future Improvements could perhaps be doubled if the ion flux were assumed to be approximately constant but the assumed full pickup ion speed were increased from the nominally assumed value of 60 to Three future steps to improve these estimates for the important O and S source rates and to explore their long-term variability may be anticipated. First, the source velocity distrikm s -1 near Io, as suggested by the Galileo PLS observations butions at Io's exobase for O and S need to be independently [Frank et al., 1996]. This would then provide an instantaneous determined in order to correct differences that may exist from ---1 x 1012 W for the pickup ion power ribbon and a total plasma torus energy input rate of at least 1.67 x 1012 W. Additional contributions to the torus energy input rate may, of course, also exist from other ion sources below the exobase (e.g., a leaky ionosphere) and from pickup of neutrals other than atomic oxygen and sulfur. Such a possibility is indeed the assumed sodium source velocity distribution. Second, the source rate of atomic sulfur for incomplete collisional cascade needs to be independently determined so as to remove the assumption that it is one half of the oxygen source rate. Third, the Voyager plasma torus description currently adopted here in the neutral cloud model may be replaced with a description suggested from recent MHD simulations of Io's interaction of the plasma torus for the Galileo epoch when it becomes with the plasma torus for the December 1995 Galileo space- available so as to explore the effects of long-term variability. craft encounter epoch [Combi et al., 1998; Linker et al., 1998; The Galileo electron and ion densities are a factor of -- 2 times Combi et al., 1999]. For these MHD simulations the total ion pickup rate near Io has been calculated to be in the range larger [Frank et al., 1996; Gurnett et al., 1996], so that the greater oxygen destruction rate from Io will be somewhat bal kg s-l, with the low value obtained for an assumed anced by the greater emission rate with the outcome largely Io intrinsic magnetic field [Linker et al., 1998] and the most recent value of 1500 kg s -1 [Combi et al., 1999] obtained for no Io intrinsic magnetic field but for an improved physical description of the pickup ions in the MHD equations. For these MHD calculations the charge exchange processes produced about 2/3 to 3/4 of the total ion pickup rate. From Table 3 the corresponding near-io estimate of 459 kg s -1 is obtained by dependent upon the electron temperature and the location of the emitting neutrals relative to Io's location. Careful modeling of the new HST/STIS simultaneous observations for the O and S extended emission data [Roesler et al., 1999; Wolven et al., 1999] and the neutral sulfur cloud measurements by Durrance et al. [1983, 1995], currently being pursued, may provide a first opportunity to address the first two improvements. adding the two higher-velocity pickup rates ( = 387 kg s -1) and the Io Lagrange-sphere portion of the incomplete Acknowledgments. We wish to thank N. Thomas for helpful discollisional cascade rate in the "outer source region" of 72 kg cussions regarding his ground-based oxygen observations for Io and s -1 (---26% of 277 kg s -1 as given by Smyth and Marconi the two reviewers for constructive comments. This work was supported [1999]). This comparison suggests that it is likely that addi- by the National Aeronautics and Space Administration through contional ion sources exist and/or that a larger scaling factor may tract NASW from the Planetary Atmospheres Program. Janet G. Luhmann thanks Robert E. Johnson and another referee be appropriate in the above estimates. for their assistance in evaluating this paper. Hence, with refined future estimates it is plausible that the total torus energy input rate introduced by neutrals from Io might be able to supply the energy budget for the plasma torus References with a ratio of the torus power input for the "outer source Ballester, G. E., Ultraviolet observations of the atmosphere of Io and region" (neutral extended clouds) to "inner source region" the plasma torus, Ph.D. thesis, Johns Hopkins Univ., Baltimore, (charge exchange processes near and below Io's exobase) likely Md., Ballester, G. E., H. W. Moos, P. D. Feldman, D. F. Strobel, and M. E. in the range of 1:1-1:3. Within this framework it is particularly Summers, Detection of neutral oxygen and sulfur emissions near Io important to note that the power ribbon charge exchange using IUE, Astrophys. J., 319, L33-L38, pickup source at Io in the "inner source region" is indeed the Ballester, G. E., J. Clarke, J. Trauger, K. Stapelfeldt, D. Crisp, heterogeneous energy source favored by Shemansky [1988] to D. Strobel, M. A. McGrath, N. Schneider, J. Ajello, and M. Combi,

10 7792 SMYTH AND MARCONI: IO'S OXYGEN SOURCE AND TORUS IMPLICATIONS Ultraviolet observations of Io with HST: WFPC2 imaging and Sandel, B. R., and A. L. Broadfoot, Io's hot plasma torus-a synoptic GHRS and FOS spectroscopy (abstract), Bull. Am. Astron. Soc., 26, view from Voyager, J. Geophys. Res., 87; , , Scherb, F., and W. H. Smyth, Variability of [O I] emission near Barbosa, D. D., F. V. Coroniti, and A. Eviatar, Coulomb thermal Io, J. Geophys. Res., 98, 18,729-18,736, properties and stability of the plasma torus, Astrophys. J., 274, 429- Scherb, F., R. J. Oliversen, M. E. Freed, W. H. Smyth, J. Corliss, R. C. 442, Woodward, P. Morgenthaler, and O. O. Lupie, Ground-based obo Brown, R. A., The Jupiter hot plasma torus: Observed electron tem- servations of [O I] 6300 Angstrom emission from Io, paper preperature and energy flow, Astrophys. J., 244, , sented at Magnetospheres of the Outer Planets Meeting, Cons. Sci. Clarke, J. T., J. Ajello, J. Luhmann, N. Schneider, and I. Kanik, de l'obs. de Paris, Paris, Aug. 9-14, Hubble Space Telescope UV spectral observations of Io passing into Schneider, N.M., and J. T. Trauger, The structure of the Io torus, eclipse, J. Geophys. Res., 99, , Astrophys. J., 450, , Combi, M. R., K. Kabin, T. I. Gombosi, D. L. DeZeeuw, and K. G. Shemansky, D. E., Energy branching in the Io plasma torus: The Powell, Io's plasma environment during the Galileo flyby: Global failure of the neutral cloud theory, J. Geophys. Res., 93, , three-dimensional MHD modeling with adaptive mesh refinement, J. Geophys. Res., 103, , Skinner, T. E., and S. T. Durrance, Neutral oxygen and sulfur densities Combi, M. R., K. Kabin, T. I. Gombosi, D. L. DeZeeuw, and K. G. in the Io torus, Astrophys. J., 310, , Powell, Interactions of Jupiter's plasma torus with the Galilean Smith, R. A., F. Bagenal, A. F. Cheng, and D. F. Strobel, On the satellites: Io and Europa, paper presented at Magnetospheres of the energy crisis in the plasma torus, Geophys. Res. Lett., 15, , Outer Planets Meeting, Cons. Sci. de l'obs. de Paris, Paris, Aug , Smyth, W. H., Neutral cloud distribution in the Jovian system, Adv. Durrance, S. T., P. D. Feldman, and H. A. Weaver, Rocket detection Space Res., 12, , of ultraviolet emission from neutral oxygen and sulfur in the Io Smyth, W. H., Energy escape rate of neutrals from Io and the implitorus, Ap. J. Lett., 267, L125-L129, cations for local magnetospheric interactions, J. Geophys. Res., 103, Durrance, S. T., P. D. Feldman, W. P. Blair, A. F. Davidsen, G. A. 11,941-11,950, Kriss, J. W. Kruk, K. S. Long, and H. W. Moos, Neutral sulfur Smyth, W. H., and M. R. Combi, A general model for Io's neutral gas emission from the Io torus measured with the Hopkins Ultraviolet cloud, I, Mathematical description, Astrophys. J. Suppl. Ser., 66, 397- Telescope, Astrophys. J., 447, , , 1988a. Frank, L. A., W. R. Paterson, K. L. Ackerson, V. M. Vasyliunas, F. V. Smyth, W. H., and M. R. Combi, A general model for Io's neutral gas Coroniti, and S. J. Bolton, Plasma observations at Io with the Ga- cloud, II, Application to the sodium cloud, Astrophys. J., 328, 888- lileo spacecraft, Science, 274, , , 1988b. Geissler, P. E., A. S. McEwen, W. Ip, M. J. S. Belton, T. V. Johnson, Smyth, W. H., and M. R. Combi, The sodium zenocorona, J. Geophys. W. H. Smyth, and A. P. Ingersoll, Galileo imaging of atmospheric Res., 96, 22,711-22,727, emissions from Io, Science, 285, , Smyth, W. H., and M. R. Combi, Io's sodium exosphere and spatially Gurnett, D. A., W. S. Kurth, A. Roux, S. J. Bolton, and C. F. Kennel, extended cloud: A consistent flux speed distribution, Icarus, 126, Galileo plasma wave observations in the Io plasma torus and near 58-77, Io, Science, 274, , Smyth, W. H., and M. L. Marconi, An explanation for the east-west Hall, D. T., G. R. Gladstone, F. Herbert, R. Lieu, and N. Thomas, Io asymmetry of the Io plasma torus, J. Geophys. Res., 103, , torus EUV emission during the comet Shoemaker-Levy impact, Geophys. Res. Lett., 22, , Smyth, W. H., and M. L. Marconi, The spatial nature of the Iogenic Henry, R. J. W., and P. G. Burke, Scattering of electrons by C, N, O, plasma source (abstract), Eos Trans. AGU, 80(17), Spring Meet. N +, O +, and O ++, Phys. Rev. A, 178, , Suppl., S200, Khurana, K. K., M. G. Kivelson, and C. T. Russell, Interaction of Io Thomas, N., High resolution spectra of Io's neutral potassium and with its torus: Does Io have an internal magnetic field?, Geophys. oxygen clouds, Astron. Astrophys., 313, , Res. Lett., 24, , Trauger, J. T., K. R. Stapelfeldt, G. E. Ballester, J. T. Clarke, and Kivelson, M. G., K. K. Khurana, R. J. Walker, C. T. Russell, J. A. WFPC2 Science Team, HST observations of [O I] emission from Io Linker, D. J. Southwood, and C. Polariskey, A magnetic signature at in eclipse (abstract), Bull. Am. Astron. Soc., 29, 1002, Io: Initial report from the Galileo magnetometer, Science, 273, 337- Volwerk, M., M. E. Brown, A. J. Dessler, and B. R. Sandel, Evidence 340, 1996a. for short cooling time in the Io plasma torus, Geophys. Res. Lett., 24, Kivelson, M. G., K. K. Khurana, R. J. Walker, J. Warnecke, C. T , Russell, J. A. Linker, D. J. Southwood, and C. Polariskey, Io's in- Wiese, W. L., M. W. Smith, and M. Glennonb, Atomic Transition teraction with the plasma torus: Galileo magnetometereport, Sci- Probabilities, Volume I, Hydrogen Through Neon (A Critical Data ence, 274, , 1996b. Comparison), National Standard Reference Data Service--National Linker, J. A., K. K. Khurana, M. G. Kivelson, and R. J. Walker, MHD Bureau of Standards 4, vol. 1, Natl. Bur. of Stand., Gaithersburg, simulations of Io's interaction with the plasma torus, J. Geophys. Md., Res., 103, 19,867-19,877, Wolven, B.C., P. D. Feldman, H. W. Moos, K. D. Retherford, D. F. Lisiecki, L., M. C. Wong, and W. H. Smyth, Io's magnetospheric Strobel, and F. L. Roesler, Emission profiles of neutral oxygen and interaction: Atmospheric heating and neutral escape (abstract), Eos sulfur in Io's exosphericorona (abstract), Eos Trans. AGU, 80(17), Trans. AGU, 80(17), Spring Meet. Suppl., S200, Spring Meet. Suppl., S201, Marconi, M. L., L. Dagum, and W. H. Smyth, Hybrid fluid/kinetic Wong, M. C., and R. E. Johnson, A three-dimensional azimuthally approach to planetary atmospheres: An example of an immediate symmetric model atmosphere for Io, 1, Photochemistry and the mass body, Astrophys. J., 496, , accumulation of a nightside atmosphere, J. Geophys. Res., 101, McGrath, M. A., and R. E. Johnson, Charge exchange cross sections 23,243-23,254, for the Io plasma torus, J. Geophys. Res., 94, , Wong, M. C., and W. H. Smyth, Model calculations for Io's atmo- Mendillo, M., J. Baumgardner, B. Flynn, and W. J. Hughes, The sphere at eastern and western elongations, Icarus, in press, extended sodium nebula of Jupiter, Nature, 348, , Moore, C. E., National Standard Reference Data Service--National M. L. Marconi, Fresh Pond Research Institute, 64 Fairfield Street, Bureau of Standards 35, vol. 1, 1H-23V, Natl. Bur. of Stand., Gaith- Cambridge, MA ersburg, Md., W. H. Smyth, Atmospheric and Environmental Research, Inc., 840 Pospieszalska, M. K., and R. E. Johnson, Monte Carlo calculations of Memorial Drive, Cambridge, MA (smyth@aer.com) plasma ion-induced sputtering of an atmosphere: SO2 ejected from Io, J. Geophys. Res., 101, , Roeslet, F. L., H. W. Moos, R. J. Oliversen, R. C. Woodward Jr., K. D. Retherford, F. Scherb, M. A. McGrath, W. H. Smyth, P. D. Feldman, and D. F. Strobel, Far-ultraviolet imaging spectroscopy of Io's (Received August 24, 1999; revised November 12, 1999; atmosphere with HST/STIS, Science, 283, , accepted November 12, 1999.)