Abstract. Following Galileo's arrival at Jupiter in fall 1995, a total of six spacecraft have now sampled the Jovian magnetosphere. Using these data sets to investigate the average location and shape of the Jovian boundaries, we fit ellipse profiles to the observations, allowing for the disk-like shape of the magnetosphere and taking account of variable solar wind pressure. We find that the subsolar magnetopause location varies with solar wind dynamic pressure to power between -1/5 and -1/4, in contrast to the terrestrial -1/6 power; this is a well-known difference attributed to the presence of hot plasma and centrifugal stretching in the Jovian magnetodisk that lessens the pressure gradients in the outer magnetosphere, resulting in its unusual responsiveness to compression. The magnetopause is less flared than the bow shock as expected, and the magnetopause shape is especially streamlined (least flared and more bullet-like) during the higher solar wind dynamic pressure conditions encountered. The average subsolar shock-to-magnetopause standoff ratio is approximately 6/5, while at low incident solar wind dynamic pressure the ratio rises to around 4/3 suggesting a blunter Earth-type magnetopause shape under these conditions. In particular, our analysis confirms that the magnetopause boundary shape is influenced by the radially inflated magnetodisk, as has been previously inferred from the stretched magnetic field lines seen within the magnetosphere. Our fits to the observations reveal that the average magnetopause boundary is indeed contracted on the north-south axis about the magnetic equator. The bow shock is not found to be so asymmetric in shape, suggesting that there is little effect of external magnetic field direction, and supporting our conclusion that the internal magnetodisk shape is the cause of the magnetopause polar flattening.
The Jovian magnetosphere is unique in our solar system. Rapid rotation together with interior mass loading by Io stretches the magnetosphere into a disk-like shape so that its equatorial cross section is quite different from that in the noon-midnight meridian. The magnetodisk wobbles with the planet's rotation period, due to the 9.6o offset between the spin and dipole axes [e.g., Smith et al., 1976; Acuna et al., 1983 and references therein]. Furthermore, this disk-like Jovian magnetosphere is more "streamlined" to flow over the poles and allows the shock to approach more closely to the magnetopause than at typical blunt solar system obstacles such as the magnetosphere of Earth. While the incident solar wind Mach number and dynamic pressure will control the bow shock strength and influence its location [e.g., Farris and Russell, 1994], the shape of the shock depends on the magnetopause obstacle shape, and the shape of the magnetopause depends on the distribution of stress within the magnetosphere. We expect the shape of the Jovian magnetopause boundary to be defined by the hot plasma and the centrifugally stretched magnetodisk that exerts a force radially outward from the axis of rotation of Jupiter. These issues have previously been addressed by models and simulations [e.g., Engle and Beard, 1980; Stahara et al., 1989; Khurana, 1997; Ogino et al., 1998]. The effect on the magnetopause shape has been inferred from features of the observed magnetospheric B field and subsolar magnetosheath [Hill et al., 1974; Lepping et al., 1981a; Thomas and Jones, 1984; Slavin et al., 1985], but until now has not been confirmed directly in observations of the Jovian magnetopause boundary at low and high magnetic latitudes. Extending the work of Huddleston et al. , we will present such evidence in this paper.
At the location of the equilibrium magnetopause, total internal pressure balances the total external pressure. The hot internal plasma in the Jovian magnetosphere (originating primarily from Io) is important in the pressure balance, resulting in an unusually large range of magnetopause standoff distances and responsiveness to solar wind flow [e.g., Lepping, 1995 and references therein]. Previous studies of Pioneer and Voyager data sets have confirmed this [Smith et al., 1978; Siscoe et al., 1980; Slavin et al., 1985]. In addition, the activity of the volcanoes of Io is episodic, possibly leading to variable mass loading within the Jovian magnetosphere and adding another source of variability to its size [e.g., Spencer and Schneider, 1996]. Typically at the magnetized planets the dynamic pressure dominates in the solar wind, the magnetic and thermal pressure dominate just outside the magnetopause boundary, and the magnetic pressure dominates within. For Jupiter, all internal plasma and magnetic pressures are important, particularly the hot particle pressure and the centrifugal force of the cold corotating plasma. While most of the mass of the Jovian middle magnetosphere is contained in the cold corotating plasma sheet [e.g., Acuna et al., 1983], the energetic particle population nevertheless dominates the internal plasma pressure [Caudal, 1986; Mauk and Krimigis, 1987], and typically just at the magnetopause the energy density of the hot plasma approximately equals that of the magnetic field [Krimigis et al., 1979; Caudal, 1986]. At times of low solar wind dynamic pressure the location of the Jovian magnetopause is greatly distended by the pressure of the hot gasses in the outer magnetosphere. We therefore expect that the shape of the boundary might also vary with its location, perhaps being less influenced by the magnetodisk shape of the middle magnetosphere during low solar wind pressure conditions. In this paper we use the available observations to explore the dependence of boundary positions on the solar wind dynamic pressure, and to investigate the shape of the magnetopause and bow shock boundaries.
2. Boundary Observations
Figure 1. Locations of the Jovian bow shock (BS, open symbols) and magnetopause (MP, solid symbols) observed by Galileo [Kivelson et al., 1997], Ulysses [Bame et al., 1992], Voyager 1 and 2 [Lepping et al., 1981b] and Pioneers 10 and 11 [Intriligator and Wolfe, 1976]. These are presented in Jupiter-centered magnetic coordinates (JSM) in both (a) XJSM , YJSM plane projection, and (b) ZJSM versus magnetic-longitude plot [from Huddleston et al., 1998].
To date, six spacecraft have visited the Jovian magnetosphere, providing multiple observations of the locations of the dynamic boundaries as displayed in Figure 1. Most recently, in November 1995 the Galileo spacecraft on its dawn-side approach to Jupiter observed multiple bow shock crossings between 215 and 130 RJ (RJ is the Jovian radius), and a single traversal through the magnetopause at 118 RJ , followed later by a possible near encounter as the spacecraft re-entered the "outer" magnetosphere probably as the result of an increase in solar wind dynamic pressure [Kivelson et al., 1997]. Shown in Figure 2 are the magnetometer data from the Galileo inbound approach, in which these bow shock and magnetopause crossings have been identified. The combined observed boundary crossings from Galileo, Ulysses, Voyagers 1, 2, and Pioneers 10, 11 provide considerable longitudinal information of these boundary surfaces (X,Y plane, Figure 1a). In the third dimension, at high ecliptic latitudes above the rotational equator there is limited coverage (notably Ulysses and Pioneer 11). However, the 9.6o tilt of the magnetic dipole axis from the spin axis provides ~19o changes in magnetic latitude on ~10-hour timescales along the spacecraft paths. Thus a spacecraft approaching Jupiter in the ecliptic plane sees periodic variations in the field components within the magnetosphere [e.g., Kivelson et al., 1997, Figures 2 and 3] as the spacecraft finds itself alternately above and below the current sheet while the planet rotates. A magnetic coordinate system can reflect this by expressing the periodically varying spacecraft location with respect to the magnetic equator.
Figure 2. Galileo magnetometer data from November 14 through 30, 1995, on approach to Jupiter [Kivelson et al., 1997]. The bow shock crossings are indicated by vertical arrows, the shading shows magnetosheath region traversals, and the single magnetopause (MP) crossing is labeled [from Huddleston et al., 1998].
On each pass, we average multiple crossings separated by <1 hour in order to reduce excessive weighting of the statistics toward times of boundary disturbance or instability. We rotate the observed crossing positions into Jupiter-centered magnetic (JSM) coordinates, where XJSM points to the Sun, ZJSM is along the projection of the Jovian dipole axis onto the X = 0 plane, and YJSM is along the projection of the magnetic equator completing the right-handed set. This coordinate system is analogous to the GSM system used for Earth. The conversion from Jupiter solar orbital (JSO) coordinates to JSM is made with reference to System III coordinates (see Dessler ) where the dipole axis is tilted toward = 201.7o in SIII(65), and SIII coordinate axes rotate with the planet.
Figure 1b illustrates the magnetic latitude versus longitude coverage of these rotated locations. A magnetic coordinate system is appropriate for our analysis because we expect the boundary shapes, particularly that of the magnetopause, to be strongly influenced by the orientation and shape of the Jovian magnetodisk. The latitudinal information in such a coordinate system has previously been neglected by those modelers who assume cylindrical symmetry when fitting the boundary profiles. (Additional refinements would also include consideration of the "hinging" of the magnetodisk and sweep-back of field lines [see Khurana, 1997], which is beyond the scope of the present paper.) Our key postulate is that for Jupiter the orientation of the magnetodisk is more important to the magnetopause boundary surface than, for example, changes in the direction of the incident solar wind.
3. Dependence of Boundary Location on Solar Wind Dynamic Pressure
Scatter in the positions of the boundary crossings is expected in response to solar wind dynamic pressure variations [e.g., Smith et al., 1978; Siscoe et al., 1980; Slavin et al., 1985]. In order to take account of these pressure changes, we require observations or estimations of the external solar wind conditions during each of the spacecraft encounters. Simultaneous observations by Voyagers 1 and 2 (0.5 AU apart) and Pioneers 10 and 11 (1.5 to 2 AU apart) allow solar wind pressure PSW at Jupiter to be inferred with reasonable or high confidence for periods while one of each spacecraft pair traversed the magnetosphere. The observations from the nearby spacecraft in the solar wind are projected to Jupiter to provide a proxy for the external conditions there. Density estimates are projected assuming 1/r 2 falloff with distance from the Sun, while solar wind speed is assumed to remain constant. The example in Figure 3 compares the projected measurements from the Voyager spacecraft for the time of Voyager 1's Jovian encounter, showing remarkably good agreement between the two data sets in the solar wind. For Ulysses and Galileo (with no nearby spacecraft), we estimate solar wind conditions based on near-Earth observations. Time of arrival is predicted based on solar wind speed and the angular separation of Earth and Jupiter in their orbits about the Sun assuming a source region on the Sun emits similar plasma for at least a partial solar rotation. We find that these are low-confidence estimates (only large-scale variations survive). Therefore, for both Galileo and Ulysses we assume the average of the projected pressure estimates for the encounter period, 0.1 nPa (and we therefore do not normalize these crossings in the profile fittings of the next section).
Figure 3. Comparison of solar wind data from both Voyager spacecraft, surrounding the time of the Voyager 1 Jupiter encounter. The Voyager 2 time axis has been shifted by a little over a day to account for the solar wind transit time over the spacecraft separation distance of ~0.5 AU, and a 1/r 2 density falloff with heliocentric distance is assumed. The data are remarkably alike, so that the Voyager 2 measurements may be taken as a proxy for the solar wind conditions incident to Jupiter while Voyager 1 was within the Jovian bow shock.
Figure 4. The extrapolated subsolar standoff distance of the Jovian magnetopause versus solar wind dynamic pressure for the Voyagers 1 and 2 observations at low magnetic latitudes.
To investigate the dependence of the magnetopause subsolar standoff distance RN (at the "nose" of the magnetopause) upon the solar wind dynamic pressure, we first must calculate a nose location for each of the observations (which occur at various positions in longitude, including many observations on the flank, as seen in Figure 1). We do this using a preliminary ellipse profile fitting to the "raw" observations (not normalized for pressure). Figure 4 shows the resulting solar wind pressure dependence of RN for the Jovian magnetopause crossings of Voyagers 1 and 2, for which we have high confidence in the external pressure estimates. The plot includes only those crossings at low magnetic latitudes (ZJSM < 20 RJ ) to exclude the effect of magnetodisk polar flattening and wobble. There is still considerable scatter in these observations. Other factors affecting the boundary locations include the amount of hot internal plasma in the magnetosphere, and whether or not the boundary was at an equilibrium position or whether it was moving when observed. We believe the latter is particularly important for the bow shock, for which there is a very large scatter, especially on the flanks. It is possible that the effects of pressure pulses from the solar wind may be transmitted from the nose and may even have a temporary effect on the boundary shape before an equilibrium position is achieved.
The magnetopause RN in Figure 4 depends on PSW to the power -0.22 (+/- 0.04), i.e., between power -1/5 and -1/4. This result is in good agreement with that of Slavin et al. , who examined Pioneer and Voyager dayside observations (after considerable averaging over those boundaries separated by <10 hours). Slavin et al. found a PSW dependence to the power -0.23 for the magnetopause standoff using magnetic pressure estimates just within the magnetosphere to infer upstream dynamic pressure, and a power -0.25 dependence for the bow shock using upstream solar wind pressure observations from just outside the shock.
Our observational result may also be compared with the recent three-dimensional (3-D) magnetohydrodynamic sinulations of a rapidly corotating Jovian magnetosphere by Ogino et al. , who ran calculations for two different incident solar wind dynamic pressures. We include their results in Figure 4 (crosses), which shows that the simulated magnetosphere is a little inflated beyond the observed but the slope of the dependence on PSW is almost the same (-0.21). In comparison to the blunt terrestrial case [Choe et al., 1973; Spreiter et al., 1966; Sibeck et al., 1991], for which the magnetopause standoff varies with PSW to the power -0.167 (i.e., -1/6), our results and those of Slavin et al.  and Ogino et al.  are consistent with the inflated corotating magnetodisk configuration and the presence of hot plasma in the outer magnetosphere, which make the Jovian system highly responsive to external pressure changes.
4. Shape of the Boundaries
The magnetopause and bow shock shape are dependent on the shape and nature of the obstacle (ionosphere / magnetosphere). To fit the observations, we first use an empirical relationship for RN (PSW) of the form obtained by Slavin et al.  to scale the observed boundary crossing positions to the average solar wind dynamic pressure PAV (~0.1 nPa). This is achieved by projecting the RN from each observation using a preliminary profile shape (fit to the raw unnormalized observations), scaling RN,obs x (PSW /PAV)0.25 for the bow shock, and RN,obs x (PSW /PAV)0.22 for the magnetopause (see previous section), and then reprojecting Robs() from the scaled RN . (We note here that the normalization technique can provide an "average" but not perfect correction for external pressure since we will find below that the best fit ellipse profile shapes are themselves PSW dependent as well as magnetic-latitude dependent.) Next, ellipse profiles for boundary standoff from Jupiter, R = L /(1+ cos), are fitted to the pressure-normalized observations in two different latitudinal slices. These are (1) the magnetic-equatorial region, | ZJSM | < 30 RJ for the bow shock, < 20 RJ for the magnetopause, and (2) at "high" latitudes, |ZJSM| >30 RJ for the bow shock, >20 RJ for the magnetopause. Since there are no observations at truely high latitudes, our classification more accurately corresponds to low and intermediate or midlatitude slices. We assume symmetry about the XJSM , YJSM plane (magnetic equator), and about the XJSM , ZJSM plane, using |YJSM| only (i.e., assuming dawn/dusk symmetry, since there are insufficient dusk-side observations to constrain the profile there).
Comparisons of the magnetic-equatorial and midlatitude fits are presented for the bow shock in Figure 5 and for the magnetopause in Figure 6. In each case, we obtain "global" fits to all points (including those on the flank and along the tail), and in addition, we obtain fits to those "dayside" points near the nose and dawn/dusk terminator. In general, the "dayside" fits are better at the nose. For the magnetic-equatorial magnetopause, because of the large number of points scattered near the subsolar region, the weighting of the few near-terminator points had to be increased proportionally in order to constrain the dayside fits, and the additional dayside fit shown for this case was obtained neglecting the Ulysses observations (that cannot be accurately normalized as discussed above, and contribute considerable scatter near the nose). The intercomparison between these different fits gives an idea of the level of uncertainty in this analysis.
Figure 5. Cross-sectional fits to the Jovian bow shock crossings (normalized to average PSW) using data from all spacecraft, for (top) low and (bottom) middle magnetic latitude region cross sections separately. In each case, both "global" (solid line) and "dayside" (dashed line) boundary profiles are fitted.
Figure 6. Cross-sectional fits to the pressure-normalized Jovian magnetopause crossings for magnetic-equatorial and midlatitude regions, in similar format to that of Figure 5.
From Figures 5 and 6, it is clear that the bow shock profile is more flared than the magnetopause. The shock-to-magnetopause standoff ratio at the subsolar point (RN at YJSM = 0 in the magnetic-equatorial profiles) is 1.12, 1.28 for the global, dayside fits, respectively, i.e., a ratio of approximately 6/5. These results are summarized in Table 1. Thus, in agreement with previous works, we find that the bow shock forms closer to the magnetopause than in Earth's case where the ratio is 1.3 to 1.4 [Spreiter et al., 1966; Slavin et al., 1985; Farris and Russell, 1994] for typical incident solar wind conditions. This is to be expected; while the Earth's magnetosphere provides a rather blunt obstacle, a more streamlined magnetodisk/magnetopause allows the shocked plasma flow to be deflected more easily over the Jovian poles, allowing the bow shock to approach the magnetopause more closely at the nose. Similar inferences have previously been made based on subsolar magnetosheath widths (from fits to Pioneer and Voyager observations) and gasdynamic model results [Stahara et al., 1989; Slavin et al., 1985]. In addition, the Mach number of the Jovian bow shock is much larger than the terrestrial Mach number, allowing greater compression of the plasma and hence a shock that can stand closer to the obstacle.
Figure 7. Comparison of the magnetic-equatorial and middle magnetic latitude cross-sectional fits for the Jovian magnetopause and bow shock boundaries. The solid, dashed, and dot-dashed line boundary profiles correspond to those of Figures 5 and 6. At the midlatitude cross sections, average <| ZJSM |> of the magnetopause observations is 45 RJ , and average <| ZJSM |> = 53 RJ for the bow shock case. The retracted location of the midlatitude magnetopause compared to its low-latitude standoff is evidence for magnetopause polar flattening (see text, and also Table 2). This is in contrast to the bow shock comparison.
In order to confirm the expected magnetopause polar flattening, Figure 7 presents a comparison of all the low and middle magnetic latitude boundary cross-section fits from Figures 5 and 6. The midlatitude magnetopause fit includes boundary crossings observed between 20 < | ZJSM | < 90 for which the average location is ZJSM = 45 RJ . The bow shock midlatitude observations are between 30 < | ZJSM | < 80 with average magnetic latitude at ZJSM = 53 RJ . For the magnetopause in Figure 7a, the comparisons show that the boundary cross section at < ZJSM > = 45 RJ above the subsolar nose stands off with XJSM = 42-46 RJ (at YJSM = 0), while at the same |YJSM| = 45 RJ dawn or duskward from the subsolar point (thin horizontal line in Figure 7) the XJSM standoff on the equatorial cross section (ZJSM ~ 0) is out at ~55 RJ . These values are tabulated in Table 2. Thus our results reveal the "squashing" of the magnetopause about the magnetic equator. In contrast, the bow shock in Figure 7b does not show such polar flattening. One might expect some effect on the bow shock due to the magnetopause obstacle shape; while there is considerable uncertainly in these fits, the flattening is clearly greater for the magnetopause. The results confirm that the internal influence of the magnetospheric configuration for the magnetopause is greater than the effect of the draped BIMF direction. The 3-D MHD simulations of Ogino et al.  are consistent with our findings for the magnetopause. Adding a strong corotation in their simulations increased the equatorial magnetopause standoff above the value at the poles by 25% at dawn and 5% at dusk, and to a lesser extent on average, the simulated bow shock is inflated at dawn and dusk by 17% and 10% above polar values (for PSW = 0.18 nPa used in the simulation). The dawn-dusk asymmetry is due to the direction of corotation [Ogino et al., 1998], with dawnside standoff greater than duskside standoff. Note that these values quoted from the simulation are at 90o from the subsolar point, while the observations we have discussed extend to the flanks in YJSM but do not extend so high in latitude. A constrained, fully 3-D boundary surface fitting to the observations is not possible because there are no observations (no constraint) in the regions directly over the poles (near XJSM = 0, YJSM = 0). Our cross-sectional fittings in Figures 5 and 6 make full use of the available data, and the comparisons near the nose (Figure 7 and Table 2) are revealing.
Figure 8. Jovian magnetopause fitted points in the Y,Z-plane view from the nose. Crosses mark the points where the low and midlatitude fits intersect the planes XJSM = 0 and XJSM = 44 RJ . The dashed-line ellipses are drawn through the points to guide the eye.
In Figure 8 we illustrate the significance of the magnetopause fitted profiles on the dayside. We show two Y,Z-plane cuts by plotting the points where the midlatitude and magnetic-equatorial fits intersect the planes XJSM = 0 and XJSM = 44 RJ . (These points are also included in Table 2.) The dashed lines are ellipse sections that connect these fitted points in each plane. The XJSM = 44 RJ plane was chosen because it contains the most sunward standoff point reached by the midlatitude fits (Figure 6, lower panel) on the Y axis. The observations determining this standoff are closely clustered, both in the X,Y plane, and also in latitude (Figure 1b) near the average |ZJSM | = 45 RJ for this fit. Thus the pair of points at YJSM = 0, ZJSM = +/- 45 RJ in Figure 8 is well defined. We are not so confident in the points at ZJSM = +/- 45, YJSM = +/- 78 RJ because of the large scatter in the observations near XJSM = 0 for the midlatitude fits, and thus the X=0-plane ellipse sections are not so reliable. Nevertheless, Figure 8 clearly illustrates the level of magnetopause flattening near the nose on the dayside magnetopause.
Figure 9. Fits to subsets of (top) high and (bottom) low incident dynamic pressure locations of the Jovian bow shock observed by the Voyager and Pioneer spacecraft.
Figure 10. Comparison of profiles of the Jovian magnetopause from combined Voyager and Pioneer observations for (top) high and (bottom) low solar wind dynamic pressure conditions.
The analysis of pressure-normalized locations in Figures 5 and 6 cannot totally eliminate effects of PSW ; there is considerable scatter in the RN versus PSW dependence (Figure 4). Possible observations of the boundary while it is moving (if there is a delay in response to PSW) contributes to this scatter. We expect that global motions of the entire magnetosphere and deflections with the direction of the incident solar wind flow should not bias the axial symmetry of the fits. No significant polar flattening is seen for the bow shock, although observations show an even greater scatter in its location. We are confident that the magnetopause flattening is real, and the bow shock is less asymmetrical, at least on the dayside near the nose. However, the boundary shapes themselves may depend on PSW . Unfortunately, the number of boundary observations is not sufficient to allow the separation of both the mid and low latitude data sets further into both low and high PSW subsets. Nevertheless, in Figures 9 and 10, we have split the "raw" (nonnormalized) bow shock and magnetopause observations into low PSW (< 0.08 nPa) and high PSW (> 0.1 nPa) sets, using only the Voyager 1 and 2 observations and Pioneer observations for which we have reasonable confidence in the corresponding pressure estimates. The best fit boundary profiles are shown in each case, and these fits are compared in Figure 11. At higher PSW , the magnetopause profile is apparently considerably more streamlined (less flared) than at times of low PSW . (Note that the low-PSW bow shock is not well constrained on the flank because of the lack of observations at XJSM < -200.) Subsolar standoff values from these fits are included in Table 1. At the nose, during low-pressure conditions (average PSW ~0.05 nPa for the low-pressure data sets) the subsolar shock-to-magnetopause standoff ratio is around 4/3, similar to that of a more blunt Earth-type configuration. (The high PSW fits cannot be compared quantitatively in this way because of the different average pressures.) At times when the magnetosphere is compressed by high solar wind dynamic pressure, one might expect a more disk-like shape at the magnetopause than at times when the pressure of the hot internal plasma may cause "ballooning" of the outer magnetosphere during low incident PSW conditions. Our results are consistent with this hypothesis.
Figure 11. Comparison of the low and high PSW bow shock (BS) and magnetopause (MP) fits obtained in Figures 9 and 10.
We have investigated the location and shape of the Jovian boundaries using the Pioneers, Voyagers, Ulysses, and Galileo spacecraft observations in a magnetic coordinate system, allowing for variable solar wind pressure and the disk-like shape of the magnetosphere. There is a large scatter in the observed crossing positions. The location and shape (and the number of crossings) of the bow shock and magnetopause boundaries observed by the spacecraft depend upon the internal magnetospheric plasma content (Io activity), the magnetodisk shape and wobble with Jovian rotation (period ~10 hours), and solar wind control by dynamic pressure variations, solar wind direction changes, and transients. The dependence of Jovian subsolar magnetopause standoff on solar wind pressure to power between -1/5 and -1/4 in agreement with results of previous studies [e.g., Slavin et al., 1985] confirms the high sensitivity of the Jovian magnetosphere to compression, compared with the -1/6 power dependence at Earth.
The bow shock is more flared than the magnetopause, as expected, and the average subsolar shock-to-magnetopause standoff ratio from our fits is ~6/5, consistent with a more streamlined obstacle for which the shock forms closer to the magnetopause than in the blunt Earth case. At the higher PSW conditions encountered, the magnetopause is apparently more streamlined (less flared) than at times of low PSW . The low-PSW subsolar shock-to-magnetopause standoff ratio of 4/3 is closer to blunt Earth-type values. This is not surprising; when the magnetosphere is greatly distended (due to the presence of hot intermal plasma) during times of low external pressure, one might expect that the magnetopause is less sensitive to the magnetodisk shape of the middle magnetosphere.
Fitting the observations in a magnetic coordinate system reveals a Jovian magnetopause surface which is flattened in the ZJSM direction, i.e., "squashed" toward the magnetic equator. This important new result confirms previous interpretations of magnetodisk features, magnetosheath widths, and model results [e.g., Engle and Beard, 1980; Slavin et al., 1985; Ogino et al., 1998]. The bow shock in our analysis does not exhibit polar flattening (although the large scatter in the observations leaves some uncertainty in the shape of the fits). This demonstrates that the internal influence of the magnetodisk is more important than the effects of the solar wind B field direction and draping in determining asymmetries of the Jovian boundaries.
Acknowledgments. Many thanks to the Galileo magnetometer team and to all those involved with the success of the Galileo project. Solar wind plasma data were obtained from NSSDC (OMNIWeb and COHOWeb) and MIT over the internet. D.E.H. thanks J. W. Belcher for his assistance in obtaining Voyager 2 data, and S. Joy for help with data from the PDS archives. This work was supported by the Jet Propulsion Laboratory under research grants JPL 958510 and JPL 958694. This work was presented at the "Magnetospheres of the Outer Planets" meeting in Boulder, Colorado, March 1997. A preliminary version was presented at the COSPAR meeting in Birmingham, England, 1996. Figures 1, 2 (and 4, modified) reprinted from Advances in Space Research, vol. 21, Huddleston et al., "The location of the Jovian bow shock and magnetopause: Galileo initial results," pp. 1463-1467, Copyright 1998, with permission from Elsevier Science. UCLA-IGPP publication 4948.
Acuna, M.H., K.W. Behannon, and J.E.P. Connerney, Jupiter's magnetic field and magnetosphere, in Physics of the Jovian Magnetosphere, edited by A.J. Dessler, pp. 1-50, Cambridge Univ. Press, New York, 1983.
Bame, S.J., et al., Jupiter's magnetosphere: Plasma description from the Ulysses flyby, Science, 257, 1539, 1992.
Caudal, G., A self-consistent model of Jupiter's magnetodisc including the effects of centrifugal force and pressure, J. Geophys. Res., 91, 4201, 1986.
Choe, J.T., D.B. Beard, and E.C. Sullivan, Precise calculation of the mangetosphere surface for a tilted dipole, Planet. Space Sci., 21, 485, 1973.
Dessler, A.J., Appendix B. Coordinate systems, in Physics of the Jovian Magnetosphere, edited by A.J. Dessler, pp. 498-504, Cambridge Univ. Press, New York, 1983.
Engle, I.M., and D.B. Beard, Idealized Jovian magnetosphere shape and field, J. Geophys. Res., 85, 579, 1980.
Farris, M.H., and C.T. Russell, Determining the standoff distance of the bow shock: Mach number dependence and use of models, J. Geophys. Res., 99, 17,681, 1994.
Hill, T.W., A.J. Dessler, and F.C. Michel, Configuration of the Jovian magnetosphere, Geophys. Res. Lett., 1, 3, 1974.
Huddleston, D.E., C.T. Russell, M.G. Kivelson, K.K. Khurana, and L. Bennett, The location of the Jovian bow shock and magnetopause: Galileo initial results, Proceedings of COSPAR'96, Adv. Space. Res., 21(11), 1463, 1998.
Intriligator, D.S., and J.H. Wolfe, Results of the plasma analyzer experi- ment on Pioneers 10 and 11, in Jupiter, edited by T. Gehrels, p. 848, Univ. of Ariz. Press, Tucson, 1976.
Khurana, K.K., Euler potential models of Jupiter's magnetospheric field, J. Geophys. Res., 102, 11,295, 1997.
Kivelson, M.G., et al., Galileo at Jupiter: Changing states of the magneto- sphere and first looks at Io and Ganymede, Adv. Space Res., 20(2), 193, 1997.
Krimigis, S.M., et al., Hot plasma environment at Jupiter: Voyager 2 results, Science, 206, 977, 1979.
Lepping, R.P., Characteristics of the magnetopauses of the magnetized planets, in Physics of the Magnetopause, Geophys. Monogr. Ser., vol. 90, edited by P. Song et al., pp. 61-70, 1995.
Lepping, R.P., L.F. Burlaga, L.W. Klein, J.M. Jessen, and C.C. Goodrich, Observations of the magnetic field and plasma flow in Jupiter's magnetosheath, J. Geophys. Res., 86, 8141, 1981a.
Lepping, R.P., M.J. Silverstein, and N.F. Ness, Magnetic field measure- ments at Jupiter by Voyagers 1 and 2: Daily plots of 48 second averages, NASA Tech. Memo., 83864, 1981b.
Mauk, B.H., and S.M. Krimigis, Radial force balance within Jupiter's dayside magnetopause, J. Geophys. Res., 92, 9931, 1987.
Ogino, T., R.J. Walker, and M.G. Kivelson, A global magnetohydro- dynamic simulation of the Jovian magnetosphere, J. Geophys. Res., 103, 225, 1998.
Sibeck, D.G., R.E. Lopez, and E.C. Roelof, Solar wind control of the magnetopause shape, location, and motion, J. Geophys. Res., 96, 5489, 1991.
Siscoe, G.L., N.U. Crooker, and J.W. Belcher, Sunward flow in Jupiter's magnetosheath, Geophys. Res. Lett., 7, 25, 1980.
Slavin, J.A., E.J. Smith, J.R. Spreiter, and S.S. Stahara, Solar wind flow about the outer planets: Gas dynamic modeling of Jupiter and Saturn bow shocks, J. Geophys. Res., 90, 6275, 1985.
Smith E.J., L.R. Davis Jr., and D.E. Jones, Jupiter's magnetic field and magnetosphere, in Jupiter, edited by T. Gehrels, p. 788, Univ. of Ariz. Press, Tucson, 1976.
Smith, E.J., R.W. Fillius, and J.H. Wolfe, Compression of Jupiter's magnetosphere by the solar wind, J. Geophys. Res., 83, 4733, 1978.
Spencer, J.R., and N.M Schneider, Io on the eve of the Galileo mission, Annu. Rev. Earth Planet. Sci., 24, 125, 1996.
Spreiter, J.R., A.L. Summers, and A.Y. Alksne, Hydromagnetic flow around the magnetosphere, Planet. Space Sci., 14, 223, 1966.
Stahara, S.S., R.R. Rachiele, J.R. Spreiter, and J.A. Slavin, A three dimensional gasdynamic model for the solar wind flow past non- axisymmetric magnetospheres: Application to Jupiter and Saturn, J. Geophys. Res., 94, 13,353, 1989.
Thomas, B.T., and D.E. Jones, Modeling Jupiter's magnetospheric currents using Pioneer data: Evidence for a low-latitude cusp, J. Geophys. Res., 89, 6663, 1984.
L. Bennett, D. E. Huddleston, K. K. Khurana, M. G. Kivelson, and C. T. Russell, Institute of Geophysics and Planetary Physics, University of California, Box 951567, 405 Hilgard Ave, Los Angeles, CA 90095-1567. (e-mail: firstname.lastname@example.org)
(Received July 31, 1997; revised January 15, 1998;
accepted January 30, 1998.)